WO2008007760A1 - Dispositif de codage et de décodage, procédé de codage et procédé de décodage - Google Patents
Dispositif de codage et de décodage, procédé de codage et procédé de décodage Download PDFInfo
- Publication number
- WO2008007760A1 WO2008007760A1 PCT/JP2007/063952 JP2007063952W WO2008007760A1 WO 2008007760 A1 WO2008007760 A1 WO 2008007760A1 JP 2007063952 W JP2007063952 W JP 2007063952W WO 2008007760 A1 WO2008007760 A1 WO 2008007760A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- coefficient
- subband
- zero
- component
- run length
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 146
- 230000015654 memory Effects 0.000 claims description 413
- 230000008569 process Effects 0.000 claims description 109
- 238000012545 processing Methods 0.000 claims description 99
- 238000000605 extraction Methods 0.000 claims description 55
- 230000010365 information processing Effects 0.000 claims description 20
- 230000006870 function Effects 0.000 claims description 5
- 230000003247 decreasing effect Effects 0.000 claims description 3
- 238000007906 compression Methods 0.000 abstract description 30
- 230000006835 compression Effects 0.000 abstract description 30
- 238000006243 chemical reaction Methods 0.000 abstract 1
- 230000000875 corresponding effect Effects 0.000 description 45
- 238000010586 diagram Methods 0.000 description 19
- 238000013500 data storage Methods 0.000 description 16
- 230000000694 effects Effects 0.000 description 10
- 238000004891 communication Methods 0.000 description 9
- 230000006854 communication Effects 0.000 description 9
- 239000000284 extract Substances 0.000 description 9
- 238000013139 quantization Methods 0.000 description 4
- 239000004065 semiconductor Substances 0.000 description 4
- 230000007423 decrease Effects 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 238000007796 conventional method Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 235000019687 Lamb Nutrition 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000007175 bidirectional communication Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000006837 decompression Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000004886 process control Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
- G06T9/007—Transform coding, e.g. discrete cosine transform
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/13—Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/18—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a set of transform coefficients
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/186—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a colour or a chrominance component
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
- H04N19/423—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation characterised by memory arrangements
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/46—Embedding additional information in the video signal during the compression process
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/63—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using sub-band based transform, e.g. wavelets
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
- H04N19/93—Run-length coding
Definitions
- Code key and decoding device code key method and decoding method
- the present invention relates to a compression / decompression technique for a two-dimensional signal, and more particularly to a technique for encoding and decoding a two-dimensional signal using wavelet transform.
- the wavelet transform is a kind of subband coding, and performs multi-level band division by repeatedly performing subband division for dividing each band in the horizontal and vertical directions on the low frequency side. Such band division is called octave division.
- octave division When sub-dividing up to 3 layers, 10 subbands as shown in Fig. 16 are obtained.
- 3LL, 3LH, 3HL, and 3HH are subbands in the lowest layer
- 2LH, 2HL, and 2HH are subbands in the upper layer
- 1LH, 1HL, and 1HH are subbands in the highest layer.
- LH is a subband with a low-pass filter in the horizontal direction and a high-pass filter in the vertical direction
- HL is a subband with a high-pass filter in the horizontal direction and a low-pass filter in the vertical direction
- HH is in the horizontal and vertical directions. Both are subbands with a high-pass filter applied.
- Each subband contains wavelet transform coefficients.
- the wavelet transform coefficient included in each subband is referred to as a subband coefficient in this specification.
- wavelet transform is performed for each color component such as YCbCr (YUV).
- YUV YCbCr
- the wavelet transform coefficients of all color components corresponding to the same layer and position are combined into one subband coefficient. Call it.
- Each color component belonging to the same subband coefficient is called a subband coefficient component.
- Wavelet transform is based on block structures such as the ability to represent a single image hierarchically by superimposing image components having multiple resolutions, JPEG block distortion, etc. It is used in JPEG2000 and other systems because it has features such as low distortion and noise. However, since the wavelet transform itself has no compression effect, it is used in combination with an entropy code method that can efficiently compress each subband coefficient.
- Non-Patent Document 1 a vector is constructed by extracting a set of coefficients corresponding to the same spatial coordinates from LH, HL, and HH subbands belonging to the same layer. Specifically, for example, in 3LH, 3HL, and 3HH, a three-dimensional beta is composed of three coefficients extracted one by one from the same spatial position. In 2LH, 2HL, and 2HH, 2 X A 12-dimensional vector is made up of 12 coefficients extracted from 4 blocks. Next, a vector with a power (magnitude) near zero is quantized to zero as an invalid vector, and all coefficients belonging to the invalid vector are excluded from the encoding target. Then, the effective vector is subjected to vector quantization using a code book corresponding to the power class of the vector.
- a Huffman code is well known as one of the entropy coding methods, and an image coding apparatus that codes each subband coefficient using the Huffman code has been proposed (for example, Patent Document 2). reference).
- Patent Document 1 Japanese Patent Application Laid-Open No. 2003-274185
- Patent Document 2 JP 2002-112039 A
- Non-Patent Document 1 Hideyasu Matsumura, Michiaki Kato, Miki Takebe, Subband of Wavelet Transform Image, Vector Quantization Method by Power Class, 1995 Image Code Symposium (PCSJ95), 1995. 10, p.121-122
- a Huffman code is added to the sign of the LH, HL, and HH subband coefficients. If used, the coefficient values of LH, HL, and HH are encoded independently, resulting in poor code efficiency. For example, in a typical image, most of the U and V coefficient values of the high-frequency HH subband are 0, and the entropy is often less than 1 bit on average. However, since the Huffman code requires a code length of at least 1 bit even if the appearance probability is very large, the code becomes redundant.
- Non-Patent Document 1 a vector is constructed by extracting a set of coefficients corresponding to the same spatial coordinates from LH, HL, and ⁇ subbands belonging to the same layer, and present in vector units.
- the method for determining validity or invalidity is suitable for a low bit rate code that assumes coarse quantization.
- the entire vector, which is a set of these coefficients is determined to be an effective vector, and all the coefficients belonging to the vector are encoded, so a good compression rate cannot be obtained. .
- An object of the present invention is to provide a code key decoding method and apparatus capable of achieving a sufficient compression rate even when a prefix code such as a Huffman code is used as the entropy code key method. It is in.
- the encoding device of the present invention extracts each component of each coefficient corresponding to the same spatial coordinate, such as LH, HL, and ⁇ subband force in the same layer in the wavelet subband space,
- the coefficient extraction means that forms a set of each of the coefficients and whether or not each component of each coefficient is significant (non-zero), and the determination result is combined for each set of extracted coefficients to generate flag information.
- the first invention of the present invention is a wavelet transform unit that divides a two-dimensional signal into subbands that are a plurality of frequency domains, and includes an LH subband, an HL subband, and a ⁇ subband belonging to the same layer. , Each component of each coefficient corresponding to the same spatial coordinate is extracted, and coefficient extracting means for forming a set of coefficients, and one or more subbands specified in advance for each set of the coefficients are specified in advance.
- a significant coefficient discriminating means for generating discriminating information for each bit by performing discriminating processing and combining the discriminating information for each set of coefficients to generate flag information of a plurality of bits, and for each set of the coefficients,
- there is a coefficient component it is characterized by having a significant coefficient sign key means for performing variable length code regardless of zero or non-zero.
- a run length determination unit that counts as a number, and a run length number encoding unit that performs variable length encoding of the run length number each time the run length number is determined, and the flag information encoding unit includes the run length number Each time the number is determined, flag information of the coefficient set whose run length is interrupted is generated and variable-length encoded, and the significant coefficient encoding means performs the determination process every time the run length number is determined.
- non-zero components are variable-length encoded, and if there are coefficient components that have not been subjected to the discrimination processing, they are variable-length encoded regardless of whether they are zero or non-zero. It is characterized by that.
- a third invention of the present invention is the above-described first invention, wherein, for each set of coefficients, the total component force of the coefficients of the LH subband and the HH subband are all of the coefficients extracted immediately before in the horizontal direction.
- a run length judging means for counting up the run length when all the components of the coefficients of the HL subband are equal to the components and equal to all the components of the coefficients extracted immediately before in the vertical direction, and the run length Each time the number is determined, a run length number code means for changing the run length number to a variable length code, and the flag information encoding means interrupts the run length each time the run length number is determined.
- the flag information of the set of coefficients is generated and variable-length code is generated, and the significant coefficient encoding means is configured to determine the coefficients and components for which the determination processing has been performed each time the run length is determined.
- the ⁇ component is variable-length encoded, and the discrimination process is performed to obtain a ⁇ If component there Ru, for they zero, and feature to be variable-length code spoon regardless nonzero.
- a fourth invention of the present invention is the above significant invention according to any one of the first to third inventions.
- the designated subband in the coefficient discriminating means includes an LH subband, an HL subband, and an HH subband.
- a fifth invention of the present invention is characterized in that, in any one of the first to fourth inventions, the designated component in the significant coefficient discriminating means includes all components constituting the coefficient.
- a sixth invention of the present invention is the invention according to any one of the first to third inventions, wherein the designated subband and component in the significant coefficient discriminating means are LH subband color difference components, HL It includes the subband color difference component and the HH subband luminance and color difference component.
- a seventh invention of the present invention is the invention according to any one of the first to third inventions, wherein the significant coefficient discriminating means is configured such that any one of a plurality of color difference components is zero in the designated subband discrimination processing. It is characterized in that 1-bit discrimination information is generated by discriminating whether or not.
- the significant coefficient discriminating unit is characterized in that both of the plurality of color difference components are inspected during the discrimination processing of the LH subband and the HL subband. It is characterized by generating discrimination information of 1 bit each by discriminating whether or not the force is.
- the significant coefficient discriminating unit determines the number of run lengths determined each time the run length is determined by the run length determination means. Discriminating information on whether or not is zero, generating discriminating information for the specified subband and the specified component at the run length start coordinate, and combining the discriminating information into flag information
- the run length encoding means performs variable length coding of the run length number only when the run length number is not zero every time the run length number is determined.
- a tenth invention of the present invention is characterized in that, in any of the first to ninth inventions, the basis function of the wavelet transform means is HAAR.
- An eleventh aspect of the present invention is a decoding device in which coefficients of a plurality of subbands obtained by wavelet transform are input as a code string, and belongs to the same layer from the code string of the subbands
- Coefficient decoding means for decoding each component of each coefficient corresponding to the same spatial coordinates of the LH subband, HL subband, and HH subband to form a set of coefficients.
- the coefficient decoding means includes flag information decoding means for decoding flag information indicating a coefficient component having a value of zero for each set of coefficients, and variable length code based on the flag information based on the non-zero coefficient.
- Significant coefficient decoding means for decoding the information, and the flag information includes one or more specified components of one or more subbands specified in advance for each set of coefficients. It includes a set of 1-bit discrimination information that discriminates whether or not the force is zero.
- run length decoding means for decoding the run length number and storing it in the run length number memory each time the coefficient decoding means performs a decoding process. If the value of the run length number memory is zero, the decoding process by the coefficient decoding means is not performed, and the value of the run length number memory is reduced and all components of all coefficients are output as zero. It is characterized by comprising a continuous length means.
- the run length number decoding means for decoding the run length number and storing it in the run length number memory each time the coefficient decoding means performs decoding processing. If the value of the run length number memory is not zero, the decoding process by the coefficient decoding means is not performed, and the value of the run length number memory is reduced, and the LH subband coefficient and the HH subband coefficient are reduced. The same length as the component decoded immediately before in the horizontal direction and the same length as the component of the coefficient extracted immediately before in the vertical direction are output as the HL subband coefficients. It is characterized by that.
- the specified subband in the flag information is an LH subband, an HL subband, and an HH subband. It is characterized by including.
- the specified component in the flag information includes all components constituting a coefficient. .
- the designated subband and component in the flag information are a color difference component of an LH subband, an HL subband, It includes color difference components, luminance of HH subband, and color difference components.
- the flag information indicates whether the plurality of color difference components are all zero in the designated subband. It includes 1-bit discrimination information that indicates
- each of the flag information indicates whether or not the plurality of color difference components are powers of zero in the LH subband and the HL subband. It includes bit discrimination information.
- the flag information includes 1-bit discrimination information indicating whether or not the run length number is zero, and the run length
- the number decoding means decodes the run length only when it is determined by the flag information that the run length is not zero.
- the basis function of the inverse wavelet transform means is HAAR.
- the twenty-first invention of the present invention includes a first step of dividing a two-dimensional signal into subbands that are a plurality of frequency domains, and an LH subband, an HL subband, and an HH subband belonging to the same layer.
- the second step of extracting each component of each coefficient corresponding to the same spatial coordinate to construct a coefficient set, and the power of one or more subbands specified in advance for each coefficient set.
- a determination process is performed to determine whether or not one or more of the specified components is zero, and 1-bit determination information is generated. The determination information is combined for each set of coefficients to have a multi-bit length.
- a twenty-third aspect of the present invention in the twenty-first aspect, for each set of coefficients, it is determined whether or not the force has all components of all coefficients being zero, and the number of consecutive occurrences of zero is linked.
- the continuous length number is Each time it is confirmed, flag information of the set of coefficients whose run length is interrupted is generated and variable-length encoded, and the fourth step is performed.
- a variable length code is added to the coefficient component that has undergone the discrimination process, and there is a coefficient component that has not undergone the discrimination process. Is characterized by variable-length codes regardless of whether they are zero or non-zero.
- the twenty-first aspect for each set of coefficients, all the components of the coefficients of the LH subband and the HH subband are the coefficients extracted immediately before in the horizontal direction.
- the run length is interrupted every time the run length number is determined.
- the non-zero component of the coefficient and component for which the discrimination processing is performed is generated every time the run length number is determined. Is encoded with variable length and the above discrimination process is performed, and there are ⁇ coefficients and components. In this case, the variable length code is used regardless of whether it is zero or non-zero.
- the twenty-fourth aspect of the present invention corresponds to the same spatial coordinates of the LH subband, the HL subband, and the HH subband belonging to the same layer from the code sequence of a plurality of subbands obtained by wavelet transform.
- a significant coefficient decoding step for decoding a variable-length code based on a non-zero coefficient component based on the flag information, the flag information for each set of coefficients. Includes a set of 1-bit discriminating information for discriminating whether one or more components of one or more sub-bands specified in advance are zero. To do.
- the second step of decoding the run length number and storing it in the run length number memory, and the value of the run length number memory is not zero
- a third step of reducing the value of the continuous length number memory and outputting all components of all coefficients as zero in place of the decoding process in the first step, there is included a third step of reducing the value of the continuous length number memory and outputting all components of all coefficients as zero.
- the run length number is decoded by decoding the run length number.
- the value of the run length number memory is not zero, instead of the decoding process in the first step, the value of the run length number memory is decreased and the LH subband
- the same value as the component decoded immediately before in the horizontal direction is output as the coefficient and the HH subband coefficient, and the same value as the component of the coefficient extracted immediately before in the vertical direction is output as the coefficient of the HL subband.
- an information processing apparatus includes a first process for dividing a two-dimensional signal into subbands that are a plurality of frequency domains, and an LH subband and an HL subband belonging to the same layer. And HH subbands, each component of each coefficient corresponding to the same spatial coordinate is extracted to form a second set of coefficients, and one or more pre-specified sub-sets for each set of coefficients. Discrimination processing is performed to determine whether one or more components of the band specified in advance are zero or not, and 1-bit discrimination information is generated, and the discrimination information is combined for each set of coefficients to generate multiple bits.
- the information processing apparatus determines, for each set of coefficients, whether or not all components of all coefficients are zero, A fifth process for counting the number of consecutive occurrences of the continuous length as a run length number, and a sixth process for variable length encoding of the run length number each time the run length number is determined.
- a fifth process for counting the number of consecutive occurrences of the continuous length as a run length number and a sixth process for variable length encoding of the run length number each time the run length number is determined.
- flag information of the coefficient set in which the run length is interrupted is generated and variable-length encoded
- the run length number is determined. If there is a coefficient component that has not been subjected to the discrimination process, a non-zero component is variable-length-signed with respect to the coefficient component that has undergone the discrimination process. It is characterized by variable length coding.
- the information processing apparatus causes the LH subband and the total component force of the HH subband coefficients for each set of coefficients to be immediately before in the horizontal direction. Equal to all components of the extracted coefficient and the total component force of the HL subband coefficient A fifth process that counts up the run length when it is equal to all components of the coefficient extracted immediately before in the direction, and a sixth code that performs variable length coding of the run length every time the run length is determined. In the third process, whenever the number of run lengths is determined, flag information of the coefficient set for which the run length is interrupted is generated and variable-length coded.
- the processing whenever the number of run lengths is determined, there is a variable length coding of a non-zero ⁇ component for the coefficient and component subjected to the discrimination processing, and the discrimination processing is performed when there is a ⁇ coefficient and component Is characterized by variable-length codes regardless of whether they are zero or non-zero.
- an information processing apparatus uses, from a plurality of subband code strings obtained by wavelet transform, an LH subband, HL subband, and HH subband in the same space belonging to the same layer.
- a first process of decoding each component of each coefficient corresponding to the coordinates to form a coefficient set is performed, and the first process is performed for a coefficient whose value is zero for each coefficient set.
- 1-bit discrimination information for discriminating whether or not the power of one or more subbands specified in advance is zero for each of the coefficient pairs. It contains the set of.
- the information processing apparatus further includes a second process of decoding the run length number and storing it in the run length memory, and the run length memory. If the value of the long number memory is not zero, a third process of reducing the value of the continuous length number memory and outputting all components of all coefficients as zero is performed instead of the decoding process by the first process. It is characterized by the fact that
- the information processing apparatus performs a second process of decoding a run length number and storing the run length number in a run length memory, and If the value of is not zero, instead of the decoding process in the first step, the value of the run length number memory is reduced, and the LH subband coefficient and the HH subband coefficient are decoded immediately before in the horizontal direction.
- a thirty-third invention of the present invention is a recording medium readable by an information processing device, wherein the information processing device has a first medium for dividing a two-dimensional signal into subbands that are a plurality of frequency regions. Processing, a second process of extracting each component of each coefficient corresponding to the same spatial coordinate from the LH subband, HL subband, and HH subband belonging to the same layer, and forming a set of coefficients, For each coefficient group, each bit is discriminated by determining whether one or more pre-designated subbands have zero or more pre-designated components. Information is generated, the discrimination information is combined for each set of coefficients to generate flag information having a plurality of bits, and the generated flag information is variable-length encoded.
- a component is variable-length encoded, and if there are components of coefficients that have not been subjected to the discrimination process, a fourth process of performing variable-length encoding is performed on them regardless of whether they are zero or non-zero.
- a thirty-fourth aspect of the present invention is a recording medium readable by an information processing device, wherein a plurality of subband code sequences obtained by wavelet transform are stored in the information processing device in the same hierarchy.
- the first processing for decoding the respective components of the coefficients corresponding to the same spatial coordinates of the LH subband, HL subband and HH subband belonging to and forming a set of coefficients is performed.
- Flag information decoding processing that decodes flag information indicating a coefficient component having a value of zero for each set of coefficients, and decoding variable-length codes for non-zero coefficient components based on the flag information.
- FIG. 1 is a block diagram of a first embodiment of the present invention.
- FIG. 2 is a flowchart showing the operation of the first exemplary embodiment of the present invention.
- FIG. 3 is a block diagram of a second exemplary embodiment of the present invention.
- FIG. 4 is a flowchart showing the operation of the second exemplary embodiment of the present invention.
- FIG. 5 is a block diagram of a third exemplary embodiment of the present invention.
- FIG. 6 is a flowchart showing the operation of the third exemplary embodiment of the present invention.
- FIG. 7 is a block diagram of a fourth embodiment of the present invention.
- FIG. 8 is a flowchart showing the operation of the fourth exemplary embodiment of the present invention.
- FIG. 9 is a block diagram of a fifth embodiment of the present invention.
- FIG. 10 is a block diagram of a sixth embodiment of the present invention.
- FIG. 11 is a block diagram of a seventh exemplary embodiment of the present invention.
- FIG. 12 is a diagram schematically showing the operation of the first exemplary embodiment of the present invention.
- FIG. 13 is a diagram schematically showing the operation of the second exemplary embodiment of the present invention.
- FIG. 14 is a diagram schematically showing the operation of the third exemplary embodiment of the present invention.
- FIG. 15 is a diagram schematically showing the operation of the fourth exemplary embodiment of the present invention.
- FIG. 16 is an explanatory diagram of wavelet transform.
- Data processing unit Flag information decoding unit Significant coefficient decoding unit Coefficient output unit Inverse wavelet transform.
- Image output unit Image output unit Image input unit Data storage unit
- Subband memory Extraction coefficient memory Coding target coefficient memory Pre-line coefficient memory Continuous length memory 326 Flag information memory
- the first embodiment of the present invention includes an image input device 11, a data storage device 12, a data processing device 13 operated by program control, and a code output device 14. Has been.
- the data storage device 12 includes a subband memory 121, an extraction coefficient memory 122, and a flag information memory 123.
- the subband memory 121 stores subband data of a plurality of hierarchies obtained by two-dimensional wavelet transform of the image input from the image input device 11.
- the extraction coefficient memory 122 stores a set of coefficients of the same spatial coordinates of the LH, H L, and HH subbands of the same hierarchy sequentially extracted from the subband memory 121.
- the extraction order of coefficient pairs follows the scan line direction from the upper left edge force to the lower right edge of the image.
- the flag information memory 123 stores flag information indicating whether each color component of each coefficient value stored in the extraction coefficient memory 122 is significant (non-zero). In the case of this embodiment, in order to display whether each color component of each coefficient value is a significant color independently, a 1-bit flag is assigned to each color component of each coefficient value.
- the data processing device 13 includes an image input unit 131, a wavelet transform unit 132, a coefficient extraction unit 133, a significant coefficient determination unit 134, a flag information code unit 135, and a significant coefficient code unit. Including 136.
- the image input means 131 inputs image data from the image input device 11 and stores it in the subband memory 121 as a 0LL subband.
- the coefficient extraction means 133 extracts LH, HL, and HH subband forces, coefficient values corresponding to the same position and spatially the same position, and stores them in the extracted coefficient memory 122.
- YCbCr Y In an image having a plurality of color components such as uv
- coefficient values of all the color components corresponding to the same position in the same hierarchy and space are extracted.
- the significant coefficient determination unit 134 determines whether each color component of each coefficient value stored in the extracted coefficient memory 122 is significant (non-zero), and sets it as a 1-bit flag. Further, the flag information is combined with all the color components of all the coefficients and stored in the flag information memory 123 as a plurality of bits of flag information.
- the flag information coding means 135 performs variable length coding on the value of the flag information memory 123 and outputs it to the code output device 14.
- Significant coefficient sign key means 136 performs variable-length coding of only significant coefficients among the coefficients in extraction coefficient memory 122 and outputs the result to code output device 14.
- the image input means 131 receives image data from the image input device 11 and stores it in the subband memory 121 as a 0LL subband (step S 101 in FIG. 2).
- the coefficient extraction hierarchy is initialized to the lowest frequency (3LH, 3HL, 3HH in FIG. 1) (step S 103).
- the coefficient extraction means 133 extracts coefficient values corresponding to the same spatial position from the LH, HL, and HH subbands in the same layer, and stores them in the extraction coefficient memory 122 (step S105). .
- the coefficient values of all the color components are extracted.
- the significant coefficient determination unit 134 determines whether or not each color component of each coefficient value stored in the extraction coefficient memory 122 is significant (non-zero), and sets it as a 1-bit flag. .
- the flag information for all the color components of all the coefficients is combined and stored in the flag information memory 123 as multi-bit flag information (steps S106 to S110).
- the significant coefficient determination unit 134 clears the flag information memory 123 to 0 (step S106).
- the significant coefficient determination unit 134 determines whether the component m of the coefficient n is significant.
- Step S107 If it is significant, a flag bit for the component m of the coefficient n is set (Step S1 08). These processes (steps S107 to S108) are repeated while changing n and m (step SI10) until the determination of all components of all coefficients is completed (step S109).
- the flag information encoding means 135 performs variable length encoding on the value of the flag information memory 123, and outputs it to the code output device 14 (step Sl l l).
- the significant coefficient sign key means 136 performs variable-length coding of only the significant coefficient among the coefficients in the extraction coefficient memory 122, and outputs the result to the code output device 14 (steps S112 to S115).
- step S112 it is determined whether or not the flag bit for the component m of the coefficient n is set. If the flag is set, the value of the component m of the coefficient n is encoded. The code is output to the code output device 14 (step S113). These processes (steps SI12 to S113) are repeated while changing n and m (step SI15) until the determination of all the components of all the coefficients is completed (step S114).
- step S116 it is determined whether or not the above processing has been completed for all coordinates in the same layer (step S116), and if not completed, the coefficient extraction coordinates are scanned to the next coordinates (step S116).
- step S118 When processing of all coordinates in the same layer is completed, it is determined whether the processing in all layers is completed (step S118), and if completed, the coefficient extraction layer is changed to the next layer. Scan (step S119) and return to step S104.
- a plurality of zero coefficients are multidimensionalized and encoded. For this reason, since the most frequently occurring zero coefficient is multidimensionalized, the average number of bits can be reduced, and a high compression ratio can be achieved even by using a simple pre-status code such as a Huffman code.
- the description of the LL subband code is omitted, but a conventional method such as LZW can be used here. Since the ratio of the processing amount and the code amount of the LL subband to the whole is small, the effect of the present invention is still large even when the conventional method is applied.
- the flag information coding means 135 has the power of a one-pass code coding method that outputs a code every time flag information and coefficient values are obtained. Generate a static Huffman code by counting the frequency in the pass and output the code in the second pass.
- the Y, U, and V components of LH, HL, and HH are all multidimensionalized.
- the appearance probability of the zero value of LH is relatively small.
- Some components may be individually signed without performing the significant coefficient determination process.
- the bit width of the flag information memory can be reduced to 8 bits or less, and the use efficiency of the memory and bus is improved.
- the second embodiment of the present invention includes a code input device 21, a data storage device 22, a data processing device 23 operated by program control, and an image output device 24. Has been.
- the data storage device 22 includes a flag information memory 221, a coefficient memory 222, and a subband memory
- the flag information memory 221 stores the flag information decoded by the data processing device 23.
- the coefficient memory 222 stores LH, HL, H of the same spatial coordinates decoded by the data processing device 23.
- the subband memory 223 stores the subband decoded by the data processing device 23. It is assumed that the DC component of the image is stored in advance in the LL subband in the subband memory 223.
- the image memory 224 stores the image data decoded by the data processing device 23.
- the data processing device 23 includes flag information decoding means 231, significant coefficient decoding means 232, coefficient output means 233, inverse wavelet transform means 234, and image output means 235.
- the flag information decoding means 231 acquires and decodes variable length code data relating to flag information from the code input device 21, and stores the decoding result in the flag information memory 221.
- Significant coefficient decoding means 232 reads flag information from flag information memory 221. Furthermore, according to the authenticity of each bit value of the flag information, it is determined whether or not each component (Y, U, V) of the coefficient (LH, HL, HH) of the same spatial coordinate in the same hierarchy is significant. . For the coefficient component whose discrimination result is significant, variable-length code data is acquired from the code input device 21, decoded, and stored in the coefficient memory 222. The coefficient 'component whose discrimination result is insignificant is stored in the coefficient memory 222 with a value of zero.
- the coefficient value of each component of the same spatial coordinate in the same hierarchy is decoded into the coefficient memory 222.
- the coefficient output means 233 reads the coefficient value stored in the coefficient memory 222 and outputs it to the subband memory 223.
- the inverse wavelet transform means 234 reads out the LL, LH, HL, and HH subband coefficient values from the subnode memory 223, performs inverse wavelet transform, and outputs the result to the image memory 224.
- the image output means 235 reads image data from the image memory 224 and outputs it to the image output device 24.
- the coefficient decoding layer is initialized to the lowest frequency (step S201).
- the flag information decoding means 231 acquires variable length code data from the code input device 21 and decodes it, and stores the decoding result in the flag information memory 221 (step S 203).
- the significant coefficient decoding unit 232 reads flag information from the flag information memory 221.
- variable-length code data is acquired from the code input device 21, decoded, and stored in the coefficient memory 222 (step S205).
- the coefficient component for which the discrimination result is insignificant is stored as zero in the coefficient memory 222 (step S206).
- step S207 it is determined whether or not the processing of steps S204 to S206 has been completed for all the coefficient components. If not completed, the incomplete coefficient components are scanned (step S207). 208), it returns to step S204.
- the coefficient output means 233 reads the coefficient value stored in the coefficient memory 222 and outputs it to the subband memory 223 (step S209).
- step S210 it is determined whether or not the processing of steps S203 to S209 has been completed for all coordinates. If not completed, uncompleted decoded coordinates are scanned (step S211). Return to S203.
- the inverse wavelet transform means 234 reads out the coefficient values of the LL, LH, HL, and HH subbands from the subband memory 223, and performs inverse wavelet transform. The result is output to the image memory 224 (step S212).
- step S213 it is determined whether or not the processing of steps S202 to S212 has been completed for all layers. If not completed, the image memory 224 is stored in the subband memory 223.
- step S214 After storing in LL, the next hierarchy is scanned (step S214), and the process returns to step S202.
- the image output means 235 reads out the image data from the image memory 224 and outputs it to the image output device 24 (step S215). Next, the effect of the present embodiment will be described.
- This embodiment is equivalent to the data decoding apparatus encoded in the first embodiment, and has the advantage that the compression rate can be improved while simple decoding using a prefix code is possible. is there.
- the third embodiment of the present invention comprises an image input device 31, a data storage device 32, a data processing device 33 that operates under program control, and a code output device 34. Has been.
- the data storage device 32 includes a subband memory 321, an extraction coefficient memory 322, an encoding target coefficient memory 323, a previous line coefficient memory 324, a run length memory 325, and a flag information memory 326. Including.
- the subband memory 321 and the extraction coefficient memory 322 are the same as the subband memory 121 and the extraction coefficient memory 122 of FIG. 1 in the first embodiment.
- the encoding target coefficient memory 323 has a capacity capable of holding one or more coefficient values of LH, HL, and HH, and stores, for example, all zero values as initial values at the time of encoding each layer. Is done. In addition, the contents of the extraction coefficient memory 322 are copied each time the run length (the number of consecutive occurrences of zero) is determined. In the case of this embodiment, each coefficient value stored in the encoding target coefficient memory 323 is an encoding target.
- the previous line coefficient memory 324 has a capacity capable of holding one line of the HL subband of the layer to be encoded, and stores, for example, all zero values as initial values when encoding each layer. Is done. In addition, the coefficient value of HL in the extraction coefficient memory 322 is appropriately copied.
- the run length memory 325 stores a zero value as an initial value at the time of each hierarchical encoding. Also, the total component force of the LH and HH subband coefficients is equal to the total component of the coefficient extracted immediately in the horizontal direction, and the total component force of the HL subband coefficient is extracted immediately in the vertical direction. Counts up when all components of the specified coefficient are equal.
- the flag information memory 326 indicates whether or not each color component of each coefficient value stored in the encoding target coefficient memory 323 is significant (non-zero), and the run length number in the run length memory 325 is significant. Stores flag information indicating whether (non-zero) data is present. In the case of the present embodiment, the same 1-bit flag is used for the components having a correlation with the frequency of the zero value among the color components of the coefficient value. The number of bits of flag information is reduced. In the present embodiment, the U and V components of the LH and HL coefficients share a flag. In the case of this embodiment, the number of bits of flag information is 8 bits.
- the data processing device 33 includes an image input unit 331, a HAAR wavelet transform unit 332, a coefficient extraction unit 333, a run length determination unit 334, a significant coefficient determination unit 335, and a flag information code input unit 336. And a significant coefficient encoding means 337 and a run length number sign ⁇ means 338.
- the image input means 331 inputs image data from the image input device 31 and stores it in the subband memory 321 as the 0LL subband.
- HAAR wavelet transform means 332 is wavelet transform means having HAAR as a basis function.
- the coefficient extraction means 333 extracts LH, HL, and HH subband forces and coefficient values corresponding to the same layer and spatially the same position, and stores them in the extracted coefficient memory 322. At this time, in an image having a plurality of color components such as YCbCr (Y UV), coefficient values of all the color components corresponding to the same position in the same layer and space are extracted.
- YCbCr Y UV
- the run length determination means 334 includes the LH coefficient value in the extraction coefficient memory 322, the LH coefficient value in the encoding target coefficient memory 323, the HH coefficient value in the extraction coefficient memory 322, and the sign key coefficient memory 32 3
- the HH coefficient value in the extracted coefficient memory 322 and the HL coefficient value of the same X coordinate included in the previous line coefficient memory 3 24 are compared. If all of the compared coefficient values match and the run length in the run length memory 325 is less than the preset maximum value, the run length in the run length memory 325 is incremented. If the HL coefficients are different, the HL coefficient value in the extraction coefficient memory 322 is transferred to the HL coefficient of the same X coordinate in the previous line coefficient memory 324.
- the significant coefficient determination unit 335 determines whether each color component of each coefficient value stored in the encoding target coefficient memory 323 is significant (non-zero). For the Y component of the LH, HL and HH coefficients, and the U and V components of the HH coefficient, the judgment result is a 1-bit flag. For the U and V components of the LH and HL coefficients, a 1-bit flag is set if at least one of the U and V components is significant. In addition, it is determined whether or not the run length number in the run length memory 325 is significant (non-zero), and similarly a 1-bit flag is used. Further, all these flag information are combined and stored in the flag information memory 326 as multi-bit flag information.
- the flag information encoding means 336 performs variable length encoding on the value in the flag information memory 326 and outputs the result to the code output device 34.
- Significant coefficient sign means 337 performs variable-length coding of only significant coefficients among the coefficients in coding target coefficient memory 323 and outputs the result to code output device 34. Further, after the code output, the coefficient value in the extraction coefficient memory 322 is copied to the encoding target coefficient memory 323.
- the run length code sign means 338 is used when the run length number included in the run length memory 325 is non-zero.
- the run length number is variable-length encoded and output to the code output device 34. Also, after the code is output, the run length in the run length memory 325 is initialized to zero.
- the image input means 331 inputs image data from the image input device 31 and stores it as a 0LL subband in the subband memory 321 (step S301 in FIG. 6).
- the coefficient extraction hierarchy is initialized to the lowest frequency (3LH, 3HL, 3HH in FIG. 5) (step S303).
- Step S304 the run length memory 325, the encoding target coefficient memory 323, and the previous line coefficient memory 324 are zero. Initialized with a value. Also, the coefficient extraction coordinates are initialized to the origin (0, 0) (Step S304
- the coefficient extraction means 333 extracts coefficient values corresponding to the same spatial position from the LH, HL, and HH subbands of the same layer, and stores them in the extracted coefficient memory 322 (step S305).
- the coefficient values of all the color components are extracted.
- the run length determination means 334 compares the LH coefficient value in the extraction coefficient memory 322 with the LH coefficient value in the encoding target coefficient memory 323, and if they are the same, proceeds to step S307. Differently
- the run length determination means 334 compares the coefficient value in the extraction coefficient memory 322 with the coefficient value in the encoding target coefficient memory 323. If they are the same, the process proceeds to step S308. Differently,
- the run length determination means 334 compares the HL coefficient value in the extraction coefficient memory 322 with the HL coefficient value of the same X coordinate in the previous line coefficient memory 324. move on. If not, go to step S311 (step S308).
- step S310 if the run length number stored in the run length number memory 325 is less than the preset maximum value, the run length determination means 334 proceeds to step S310. If it is greater than or equal to the maximum value, proceed to step S311 (step S310).
- the run length determination means 334 increments the run length number in the run length number memory 325, and proceeds to step S332 (step S310).
- the significant coefficient determination unit 335 performs a significant coefficient determination process (steps S311 to S327).
- the significant coefficient determination means 335 performs the flag information memory 3
- the flag value in 26 is cleared to zero (step S311).
- the significant coefficient determination means 335 checks whether or not the run length number in the run length memory 325 is non-zero (step S312). If it is non-zero, the flag in the flag information memory 326 is checked. Among the bits, the bit for the run length is set (step S313).
- the significant coefficient determination means 335 performs the Y component of the LH coefficient in the encoding target coefficient memory 323. Is non-zero (step S314). If non-zero, the bit for the Y component of the LH coefficient is set out of the flag bits in the flag information memory 326 (step S315).
- the significant coefficient determination means 335 checks whether the U or V component of the LH coefficient in the sign target coefficient memory 323 is non-zero (step S316), and at least one is non-zero. If there is, the bit for the UV component of the LH coefficient is set among the flag bits in the flag information memory 326 (step S317).
- the significant coefficient determination means 335 checks whether the Y component of the HL coefficient in the encoding target coefficient memory 323 is non-zero (step S318). Among the flag bits in the register 326, the bit for the Y component of the HL coefficient is set (step S319). Also, it is checked whether the U or V component of the HL coefficient in the code target coefficient memory 323 is non-zero (step S320). If at least one is non-zero, the flag bit in the flag information memory 326 is checked. Among them, the bit for the UV component of the HL coefficient is set (step S321).
- the significant coefficient determination means 335 checks whether or not the Y component of the HH coefficient in the encoding target coefficient memory 323 is non-zero (step S322), and if it is non-zero, the flag information memory 3 Among the flag bits in 26, the bit for the Y component of the HH coefficient is set (step S323). Similarly, it is checked whether or not the U component of the HH coefficient in the sign key coefficient memory 323 is non-zero (step S324). If it is non-zero, the flag bits in the flag information memory 326 are Set bit for U component of HH coefficient (step S325).
- step S326 it is checked whether or not the V component of the HH coefficient in the encoding target coefficient memory 323 is non-zero (step S326), and if it is non-zero, the HH coefficient of the flag bits in the flag information memory 326 is checked. Set the bit for the V component (step S327).
- the flag information encoding means 336 performs variable-length encoding of the flag in the flag information memory 326 (step S328), and outputs it to the code output device 34.
- the significant coefficient signing means 337 encodes the significant coefficient in the same manner as the significant coefficient coding processing in the first embodiment shown in steps S112 to S114 in FIG. Perform (Step S329). Specifically, among the Y, U, and V components of the LH, HL, and HH coefficients, the variable length code is input in a predetermined order for the component in which the flag bit of the flag information memory 326 is set.
- the variable length code is input in a predetermined order for the component in which the flag bit of the flag information memory 326 is set.
- the run length number code key means 338 performs variable length coding on the run length number when the run length number included in the run length number memory 325 is non-zero, and outputs it to the code output device 34. (Step S330).
- the significant coefficient encoding means 337 and the run length number encoding means 338 copy the coefficient value of the extraction coefficient memory 322 to the encoding target coefficient memory 323, and initialize the run length number. Also, the HL coefficient value in the extraction coefficient memory 322 is transferred to the HL coefficient of the same X coordinate in the previous line coefficient memory 324 (step S331).
- step S332 it is determined whether or not the processing of steps S305 to S331 has been completed for all coordinates (step S332), and if not completed, uncompleted decoded coordinates are scanned (step S333). Return to S305.
- the significant coefficient encoding means 337 performs the significant coefficient determination process similar to steps S311 to S327 (step S334), and the flag information encoding means 336.
- the sign in the flag information memory 326 is encoded and output (step S335), the significant coefficient sign means 337 encodes and outputs the significant coefficient (step S336), and the run length sign sign means 338 signs and outputs the run length number (step S337).
- step S338 it is determined whether or not the processing of steps S304 to S337 has been completed for all layers (step S338), and if completed, the coefficient extraction layer is scanned to the next layer (step S338). S 339), returning to step S304.
- the flag information is 8 bits, so the memory and bus utilization efficiency can be improved.
- the U component and V component of LH and the U component and V component of HL are each integrated into a 1-bit flag.
- the flags of the U and V components may be integrated according to the HH coefficient.
- the U and V components of the HH coefficient originally have many zero values, the contribution to the compression ratio due to multidimensionalization is also large. For this reason, it is advantageous in terms of compression ratio to determine significance individually.
- the compression rate of graphic images can be greatly improved by combining the number of consecutive occurrences of these coefficients in combination with the HAAR wavelet transform.
- this continuous length determination method can generally obtain a good compression rate when applied not only to graphic images but also to natural images.
- the HL coefficient is analyzed not for the run length in the vertical direction but for the coincidence length compared to the HL coefficient value one line before.
- This slide dictionary method improves the continuity of memory access and achieves higher speed than counting vertical run length.
- the run length value is encoded by counting the LH, HL, HH coefficients, Y, U, and V components separately, they are handled in a lump, thereby simplifying the processing and reducing the compression rate. Improvement is achieved.
- the reason for this is that in many images, there is a correlation between the run length cutting positions of each component (for example, when the run length is interrupted by the Y component of LH, the rate at which the run length is interrupted by other components is also high. ). For this reason, the compression rate is unlikely to decrease even if the run length numbers are not recorded individually but are counted and entered in a lump.
- the zero value is multidimensionalized in this embodiment. Therefore, the reduction in compression rate due to the re-encoding of the zero value is negligible.
- the power at which the run length number 0 appears most frequently Whether the number is 0 or more is multi-dimensional as flag information, so the compression rate hardly decreases even when the run length number 0 occurs frequently. Also, whether the run length number is 1 or more is highly correlated with the significance of the LH, HL, and HH coefficients (for example, if the run length is all zero, the run length is 1 or higher). Since these are multidimensional, the compression rate is further improved.
- the YCbCr (YUV) color system is taken as an example, and the LH U-component and V-component, HL U-component and V-component flags combined, and other color systems are used.
- it is effective to combine components that correlate with the frequency of zero values (for example, color difference components, a * and b * in the case of L * a * b *).
- the encoding target coefficient memory initially set in each layer is first encoded in each layer.
- the flag bits of the LH, HL, and HH coefficients, excluding the run length flag bits are fixed to their initial values, and therefore redundant when output as codes. Therefore, in the initial code output at each layer, only the run length number may be variable-length encoded and output.
- a method for determining the run length when all the components of LH, HL, and HH are zero can be employed as the run length determination method.
- This run length determination method slightly improves the compression rate for natural images than the method of counting run lengths of the same value. However, the compression rate is greatly reduced for graphic images.
- the fourth embodiment of the present invention includes a code input device 41, a data storage device 42, a data processing device 43 that operates under program control, and an image output device 44. Has been.
- the data storage device 42 includes a flag information memory 421, a run length number memory 422, a previous line coefficient memory 423, a coefficient memory 424, a sub-node memory 425, and an image memory 426.
- the flag information memory 421 stores flag information decoded by the data processing device 43.
- the run length number memory 422 stores a zero value as an initial value at the time of decoding each layer, and stores the decoded run length as appropriate after that.
- the previous line coefficient memory 423 has a capacity capable of holding one line of the HL subband of the hierarchy to be decoded, and all zero values are stored as initial values at the time of decoding each hierarchy.
- the coefficient memory 424 has a capacity capable of holding one or more coefficient values of LH, HL, and HH, and all zero values are stored as initial values at the time of decoding each layer.
- the sub-band memory 425 has a capacity capable of holding only the LL sub-band, and stores a DC component of the image by force.
- the image memory 426 stores the image data decoded by the data processing device 43.
- the data processing device 43 includes a flag information decoding unit 431, a run length decoding unit 432, a run length unit 433, a significant coefficient decoding unit 434, a HAAR inverse wavelet transform unit 435, and an image output unit 436. Including.
- the flag information decoding unit 431 obtains variable length code data from the code input device 41, decodes it, and stores the decoding result in the flag information memory 421.
- the run-length number decoding means 432 acquires and decodes variable-length code data from the code input device 41 and stores the decoding result in the run-length number memory 422.
- the run length means 433 reads the run length number from the run length memory 422, and determines whether it is non-zero. If the run length number is non-zero, the run length number is decremented and the run length memory 422 is updated, and the HL coefficient value of the same X coordinate in the previous line coefficient memory 423 is changed to the HL coefficient in the coefficient memory 424. Copy.
- the significant coefficient decoding unit 434 reads flag information from the flag information memory 421. Further, it is determined whether or not each component (Y, U, V) of the coefficient (LH, HL, HH) of the same spatial coordinate in the same hierarchy is significant according to the authenticity of each bit value of the flag information. For coefficient components for which the discrimination result is significant, variable-length code data is acquired from the code input device 41, decoded, and stored in the coefficient memory 424. The coefficient component for which the discrimination result is insignificant is stored in the coefficient memory 424 with a value of zero. The U and V components of the LH and HL coefficients are integrated into a 1-bit flag. If the flag is set, both U and V components are decoded.
- the coefficient value of each component of the same spatial coordinate in the same hierarchy is stored in the coefficient memory 424. Is decrypted.
- HAAR inverse wavelet transform means 435 reads the coefficient values of the LH, HL, and HH subbands from the coefficient memory 424 and outputs the coefficient values of the LL subbands corresponding to the same spatial coordinates in the same layer from the subband memory 425. read out. Further, the HA AR inverse wavelet transform is performed on these coefficient values, and the result is output to the image memory 426.
- the image output means 436 reads image data from the image memory 426 and outputs it to the image output device 44.
- the coefficient decoding layer is initialized to the lowest frequency (step S401).
- the run length number memory 422, the coefficient memory 424, and the previous line coefficient memory 423 are initialized with zero values. Also, the coefficient extraction coordinates are initialized to the origin (0, 0) (step S402).
- the run length means 433 reads the run length number from the run length memory 422, and determines whether it is non-zero (step S403). If the run length power is ⁇ , the process proceeds to step S406. If the run length number is non-zero, the run length number is decremented and the run length memory 422 is updated (step S404), and the HL coefficient of the coefficient memory 424 is changed from the previous line coefficient memory 423 to the HL Transfer to numerical value (step S405). Then, the process proceeds to step S431.
- step S 406 flag information decoding means 431 acquires and decodes variable length code data from code input device 41, and stores the decoding result in flag information memory 421.
- the significant coefficient decoding means 434 reads flag information from the flag information memory 421 and performs significant coefficient decoding processing (steps S407 to S428).
- significant coefficient decoding means 434 checks whether the Y component flag of the LH coefficient is set (step S407). If standing, the variable length code is acquired from the code input device 41 and decoded, and the decoding result is stored in the Y component of the LH coefficient in the coefficient memory 424 (step S408). If not, zero is stored in the Y component of the LH coefficient in the coefficient memory 424 (step S409).
- the significant coefficient decoding means 434 checks whether the UV component flag of the LH coefficient is set (step S410). If standing, the variable length code is sent from the code input device 41. Are obtained in succession and decoded, and the decoding result is stored in the U and V components of the LH coefficient in the coefficient memory 424 (step S411). If not, zero is stored in the U and V components of the LH coefficient in the coefficient memory 424 (step S412).
- the significant coefficient decoding means 434 checks whether or not the Y component flag of the HL coefficient is set (step S413). If standing, the variable length code is acquired from the code input device 41 and decoded, and the decoding result is stored in the Y component of the HL coefficient in the coefficient memory 424 (step S414). If not, zero is stored in the Y component of the HL coefficient in the coefficient memory 424 (step S415).
- the significant coefficient decoding means 434 checks whether the UV component flag of the HL coefficient is set (step S416). If standing, two variable length codes are obtained in succession from the code input device 41 and decoded, and the decoding results are stored in the U and V components of the HL coefficient in the coefficient memory 424 (step S417). ). If not, zero is stored in the U and V components of the HL coefficient in the coefficient memory 424 (step S418).
- the significant coefficient decoding means 434 checks whether or not the Y component flag of the HH coefficient is set (step S419). If standing, the variable length code is acquired from the code input device 41 and decoded, and the decoding result is stored in the Y component of the HH coefficient in the coefficient memory 424 (step S420). If not, zero is stored in the Y component of the HH coefficient in the coefficient memory 424 (step S421).
- the significant coefficient decoding means 434 checks whether or not the U component flag of the HH coefficient is set (step S422). If standing, the variable length code is acquired from the code input device 41 and decoded, and the decoding result is stored in the U component of the HH coefficient in the coefficient memory 424 (step S423). If not, zero is stored in the U component of the HH coefficient in the coefficient memory 424 (step S424).
- the significant coefficient decoding means 434 checks whether or not the V component flag of the HH coefficient is set (step S425). If standing, the variable length code is acquired from the code input device 41 and decoded, and the decoding result is stored in the V component of the HH coefficient in the coefficient memory 424 (step S426). If not, zero is stored in the V component of the HH coefficient in the coefficient memory 424 (step S427). [0202] When the above processing is completed, the significant coefficient decoding means 434 transfers the HL coefficient value decoded in the coefficient memory 424 to the HL coefficient of the same X coordinate in the previous line coefficient memory 423 (step S428).
- the run length number decoding means 432 decodes the run length number and stores it in the run length number memory 422 (steps S429 and S430). Then, proceed to step S431.
- step S431 the HAAR inverse wavelet transform means 435 reads the coefficient values of the LH, HL, and HH subbands from the coefficient memory 424 and calculates the coefficient values of the LL subbands corresponding to the same spatial coordinates in the same layer. Read from the subband memory 425 (step S431). Further, the HAAR inverse wavelet transform is performed on these coefficient values, and the result is output to the image memory 426 (step S432).
- step S433 it is determined whether or not the processing in steps S403 to S432 has been completed for all coordinates. If not completed, uncompleted decoded coordinates are scanned (step S434). Return to S403.
- step S435 it is determined whether or not the processing of steps S402 to S433 has been completed for all layers. If not completed, the data stored in the image memory 426 is used as a new LL subband. It is regarded as a component and stored in the LL of the subband memory 425, then the next hierarchy is scanned (step S436), and step S402 is returned.
- the image output means 436 reads out the image data from the image memory 426 and outputs it to the image output device 44 (step S437).
- the present embodiment corresponds to a decoding apparatus for data encoded in the third embodiment.
- the flag information is 8 bits, so the memory and bus utilization efficiency can be improved. Also, the compression ratio can be improved by the run length.
- the continuity of memory access is improved by adopting a slide dictionary method that analyzes the match length compared to the HL coefficient value one line before the run length in the vertical direction. Up and speedup is achieved.
- run length value is handled in a lump rather than separately counting and coding the LH, HL, HH coefficients, Y, U, and V components, thereby simplifying the processing and improving the compression rate. Is achieved.
- HAAR is used as the basis of wavelet transform
- inverse wavelet operation can be performed immediately after decoding LH, HL, and HH coefficients of the same layer and the same spatial coordinates.
- the memory capacity that does not require the contents of the coefficient memory to be saved in the subband memory can be reduced, and high speed can be achieved by avoiding access to real memory such as DRAM.
- a decoding apparatus that decodes a code string encoded by a method of determining a continuous length when all components of LH, HL, and HH are zero is also conceivable.
- the decoding process by the flag information decoding means 431 and the significant coefficient decoding means 434 is not performed, and the run length means 433 uses the value in the run length memory 422. Decrease and output all components of all coefficients as zero.
- Fig. 9 is a block diagram of a bidirectional communication terminal incorporating a coding and decoding apparatus according to the fifth embodiment of the present invention.
- the terminal is provided with a CPU 501 as a program control processor, and is connected to the cache memory 502 and the main memory 503 via an internal bus.
- the internal bus is further connected to an external bus 504 via a port.
- the external bus 504 includes a program memory 505 storing necessary programs, a data memory 506, an interface 507 for connecting a camera 508, an interface 509 for connecting a display 510,
- the interface 511 for connecting the communication control unit 512 and the like are also connected.
- Communication control unit 512 is connected to network 513. If it is a mobile phone, the communication controller 5 Reference numeral 12 includes a radio communication unit and a channel control unit, and a network 513 is a mobile communication network.
- the program memory 505 stores an encoding program, a decoding program, a main program for controlling the overall operation of the communication terminal, and the like.
- the sign key program is expressed by the flowchart shown in any of FIGS. 2 and 6 described in the first and third embodiments, and the decoding program is expressed by the flowcharts shown in FIGS. Is.
- the above-described processes of the code key program and the decoding program are executed under process control by the main program.
- the data memory 506 includes a subband memory, a coefficient memory, and a flag information memory.
- the subband memory is, for example, the subband memory 121 in FIG. 1 or the subband memory 223 in FIG.
- the coefficient memory is, for example, the extraction coefficient memory 122 in FIG. 1 or the coefficient memory 222 in FIG.
- the flag information memory is, for example, the flag information memory 123 in FIG. 1 or the flag information memory 221 in FIG.
- the camera 508 corresponds to the image input devices 11 and 31 in the encoding device of FIGS. 1 and 5, and the display 510 corresponds to the image output devices 24 and 44 in the decoding device of FIGS. .
- the image data captured by the camera is wavelet transformed and encoded as described above, and the code string is transmitted to the counterpart terminal via the network 513.
- the code sequence of the received image data is decoded as described above, inversely wavelet transformed, and displayed on the display 510.
- the detailed operation is as described above, and any of them can use the cache memory 502 as a coefficient memory.
- the encoding device can be realized by executing each control program on the CPU as described above, but can also be realized by hardware.
- a configuration in which each process of the first embodiment described above is converted into nodeware is illustrated, but the third embodiment is similarly configured.
- FIG. 10 is a block diagram showing an example of a coding device according to the sixth embodiment of the present invention.
- Image data input by an image input device 601 such as a camera is stored in the subband memory 602, and sequentially wavelet transformed by the wavelet transform unit 603 as described above.
- the subband coefficients LL, LH, HL, and HH obtained in this way are stored in the subband memory 602.
- the coefficient extraction unit 604 extracts LH HL and HH coefficient sets of the same hierarchy and the same coordinates from the subband memory 602 and stores the coefficient sets in the coefficient register 605.
- the coefficient sets stored in the register 605 are determined by the significant coefficient determination unit 606 as to whether each is significant and stored in the flag information register 607.
- the flag information encoding unit 609 performs variable length encoding on the contents of the flag information register 607 and outputs the result to the code output unit 610.
- Significant coefficient encoding unit 608 performs variable-length coding on only significant coefficients in coefficient register 605 and outputs the result to code output unit 610.
- the decoding device can also be realized by hardware that can be realized by executing each control program on the CPU as described above.
- a configuration in which each process of the second embodiment described above is converted into nodeware is exemplified, but the same configuration is also applied to the fourth embodiment.
- FIG. 11 is a block diagram showing an example of a decoding device according to the seventh embodiment of the present invention.
- the code string sent out by the code input device according to the sixth embodiment is inputted by the code input device 701.
- the flag information decoding unit 702 decodes the variable length code and stores it in the flag information register 703.
- the significant coefficient decoding unit 704 determines whether each coefficient is significant from each bit of the flag information register 703, decodes the variable-length code according to the significant coefficient, and stores it in the coefficient register 705, which is not significant.
- the zero value is stored in the coefficient register 705 for the coefficient.
- the value of the coefficient register 705 generated in this way is stored in the subband memory 707 as a coefficient of the same hierarchy and the same spatial coordinates.
- the inverse wavelet transform unit 706 reads the LH, HL, and HH subbands from the subband memory 707, performs inverse wavelet transform, and stores the result in the subband memory 707. It is output to the finally obtained LL subband 1S image output unit 708.
- flag information is variable-length encoded, only significant coefficients are variable-length encoded.
- a prefix code such as a Huffman code
- Fig. 12 is a diagram schematically showing a sign key operation according to the first embodiment of the present invention.
- a personal computer is provided as a data processing device
- a semiconductor memory (included in the personal computer) is provided as a data storage device
- a magnetic disk device is provided as a code output device.
- the image data is given as a file on the magnetic disk device.
- image data is read from the magnetic disk device, and stored in the subband memory as the 0LL subband Z101.
- the subband Z102 of each layer divided into octaves by the wavelet transform unit is configured in the subband memory.
- the coefficient extraction means extracts LH, HL, and HH subband forces from coefficient values corresponding to the same layer and space coordinates.
- LH coefficient Z106, HL coefficient Z107, and HH coefficient Z108 are extracted and stored in the extracted coefficient memory.
- the significance coefficient judging means checks whether each color component (Y, U, V) of LH coefficient Z106, HL coefficient Z107, and HH coefficient Z108 is significant (non-zero).
- the corresponding bit flag is set in the flag information register Z109.
- the FLHY and FLHU bits are 1 and the V component is 0, so the FLHV bits are 0.
- the value of the flag information register Z109 thus obtained is 01A2h (hexadecimal notation).
- the flag information code input means converts this value into a variable length code and outputs it as a code Z111 to the file Z110.
- the significant coefficient encoding means extracts only the significant coefficient from the LH coefficient Z106, the HL coefficient Z107, and the HH coefficient Z108, performs variable length coding, and outputs it to the file Z110. For example, since Y and U of the LH coefficient Z106 are non-zero, they are output as variable length codes Z112 and Z113. Similarly, since ⁇ of HL coefficient Z107 is non-zero, it is output as variable length code Z114. U Since U of coefficient Z107 is non-zero, it is output as variable length code Z115.
- FIG. 13 is a diagram schematically showing a decoding operation according to the second embodiment of the present invention.
- a personal computer is provided as a data processing device
- a semiconductor memory (included in the personal computer) is provided as a data storage device
- a magnetic disk device is provided as an image output device.
- the code data is given as file Z201 on the magnetic disk device.
- the flag information decoding means reads the variable length data Z202 from the code data Z201, decodes it, and stores it in the flag information memory Z207.
- the decoding result of the flag information is 01A2h (hexadecimal notation).
- the significant coefficient decoding means refers to each bit of the flag information memory Z207, and when the bit is standing, considers that the coefficient value corresponding to the bit position is significant, and starts from the code data Z2 01. Variable length code is read, decoded, and stored in LH coefficient memory Z208, HL coefficient memory Z209, and HH coefficient memory Z210. If no bit is set, zero is written to the coefficient value corresponding to the bit position.
- variable length code Z203 is read and decoded to obtain the coefficient value 2, and the coefficient value 2 is stored in the Y component of the LH coefficient memory Z208.
- variable length codes Z204, Z205, and Z206 are read and decoded, and the result is the U component in the LH coefficient memory, the Y component in the HL coefficient memory, and the HH coefficient. Store in the U component of the memory. For the other components, the zero value is stored because the bit is not set.
- FIG. 14 is a diagram schematically showing a sign key operation according to the third embodiment of the present invention.
- a personal computer is provided as a data processing device
- a semiconductor memory (included in the personal computer) is provided as a data storage device
- a magnetic disk device is provided as a code output device.
- the image data is given as a file on the magnetic disk device.
- image data is read from the magnetic disk device, and the subband Z301 of each layer divided into octaves by the HAAR wavelet transform means is configured in the subband memory. Thereafter, the LH, HL, and HH coefficients of each layer are sequentially encoded. Hereinafter, the encoding of 1LH, 1HL, and 1HH will be described.
- the run length number memory Z3021 is initialized to 0, and the encoding target coefficient memory Z3022 is all initialized to 0. Also, the previous line coefficient memory Z3023 is all initialized to zero. Also, the coefficient extraction coordinates are initialized to (0, 0).
- the coefficient extraction means also extracts LH, HL, and HH subband forces from coefficient values corresponding to the same layer and space coordinates.
- the LH coefficient (3, 2, 0), HL coefficient (0, 0, 0), and HH coefficient (0, 2, 0) at the coordinates (0, 0) are extracted and stored in the extraction coefficient memory Z3024. As a result, each coefficient memory and continuous length memory are in the state of Z302.
- the significant coefficient determination means includes the coefficient value of the encoding target coefficient memory Z3022 and the run length number memory Z302. It is checked whether the value of 1 is non-zero, and the result is stored in the flag information memory Z3031. Here, since all the values are 0, the flag information is 0.
- the flag information encoding means performs variable length encoding on the value of the flag information memory Z3031.
- Significant coefficient encoding means performs variable length encoding of only significant coefficients from the encoding target coefficient memory Z3022. Here, no sign of the coefficient is generated because there is no significant coefficient.
- the run length number encoding means checks whether or not the run length number memory Z3021 is non-zero. If it is non-zero, the run length number is variable-length encoded. Here, since the run length number is 0, the run length code is not generated.
- variable length code of Z304 is generated at the coordinates (0, 0) and output to the file.
- Each coefficient value in the extraction coefficient memory Z3024 is transferred.
- LH coefficient (3, 2, 0), HL coefficient (0, 0, 0), HH coefficient (0, 2, 0) of coordinates (1, 0) are extracted and stored in extracted coefficient memory Z3054 The As a result, the coefficient memory and the run length number memory are in the Z305 state.
- the run length determination means is that the LH coefficient (3, 2, 0) and HH coefficient (0, 2, 0) are the LH coefficient (3, 2, 0) and HH coefficient of the encoding target coefficient memory Z3052.
- Check whether the force is equal to (0, 2, 0) and whether it is equal to the previous line coefficient memory Z3053 corresponding to the HL coefficient (0, 0, 0) force X coordinate 1.
- the run length number Z3051 is less than the maximum value (255), the run length condition is satisfied.
- the value in the run length memory Z3051 is incremented to 1, and the processing at the coordinate (1,0) is completed.
- LH coefficients (3, 2, 0), HL coefficients (0, 4, 0), and HH coefficients (0, 2, 0) at coordinates (4, 0) are extracted and stored in the extracted coefficient memory Z3064.
- the run length determination means is that the LH coefficient (3, 2, 0) and HH coefficient (0, 2, 0) are the LH coefficient (3, 2, 0) and HH coefficient of the encoding target coefficient memory Z3062.
- Check whether the force is equal to (0, 2, 0) and whether it is equal to the previous line coefficient memory Z3063 corresponding to the HL coefficient (0, 4, 0) force X coordinate 4.
- the significant coefficient determination means includes the coefficient value of the encoding target coefficient memory Z3062 and the run length number memory Z306.
- the most significant bit of the flag information memory Z3071 is a run length flag.
- the run length is 3.
- the second bit from the most significant bit of the flag information memory Z3071 is the Y component flag of LH.
- the Y component of H is 3 and is non-zero, so 1 is stored in the flag bit.
- the third bit from the most significant bit of the flag information memory Z3071 is a flag for the UV component of LH.
- the fourth bit from the most significant bit of the flag information memory Z3071 is the Y component flag of the HL. H
- the fifth bit from the most significant bit of the flag information memory Z3071 is a flag for the UV component of HL.
- the U and V components of HL are 0 and 0, respectively, and they are all zero, so 0 is stored in the flag bit.
- the sixth bit from the most significant bit of the flag information memory Z3071 is a flag of the Y component of HH. H
- the seventh bit from the most significant bit of the flag information memory Z3071 is a flag of the U component of HH. H
- the eighth bit from the most significant bit of the flag information memory Z3071 is a flag of the V component of HH. H
- the flag information encoding means performs variable length encoding on the value (E2h, hexadecimal notation) of the flag information memory Z3071 generated as described above.
- Significant coefficient encoding means variable length only significant coefficients from the encoding target coefficient memory Z3062. Encode. In LH and HL, when either or both of U and V components are non-zero values, both U and V components are variable length encoded. Here, Y component (3) of LH, UV component (2, 0), and U component (2) of HH are encoded.
- the run length number encoding means checks whether or not the run length number memory Z3061 is non-zero. If it is non-zero, the run length number is variable-length encoded. Here, since the run length is 3, a code with run length 3 is generated.
- variable length code of Z308 is generated at the coordinates (4, 0) and output to the file.
- each coefficient value of the extraction coefficient memory Z3064 is transferred.
- the significant coefficient determination means checks whether the coefficient value of the encoding target coefficient memory Z3092 and the value of the run length number memory Z3091 are non-zero, and stores the result in the flag information memory Z3101.
- the flag information encoding means performs variable-length encoding on the value (EAh, hexadecimal notation) of the flag information memory Z3101 generated as described above.
- the significant coefficient encoding means performs variable-length encoding of only significant coefficients from the encoding target coefficient memory Z3092.
- Y component (3) of LH, UV component (2, 0), UV component (4, 0) of HL, and U component (2) of HH are encoded.
- the run length number encoding means checks whether or not the run length number memory Z3091 is non-zero. If it is non-zero, the run length number is variable-length encoded. Here, since the run length is 4, a code with run length 4 is generated.
- Each coefficient value in the extraction coefficient memory Z3094 is transferred.
- FIG. 15 is a diagram schematically showing a decoding operation according to the fourth embodiment of the present invention.
- a personal computer is provided as a data processing device
- a semiconductor memory (included in the personal computer) is provided as a data storage device
- a magnetic disk device is provided as an image output device.
- the code data is given as a file on the magnetic disk device.
- each ILL, 1LH, 1HL, and 1HH sub-band of Z401 shows the Y, U, and V component values of each coefficient in each spatial coordinate.
- the values of the Y component, U component, and V component at the coordinates (0, 0) of the LL coefficient stored in the subband memory Z401 are a, b, and c, respectively.
- the Y, U, and V components of the LH coefficient obtained by decoding the coordinates (0, 0) are 3, 2, and 0, respectively.
- the Y, U, and V components of the HL coefficient are 0, 0. , 0, indicating that the Y, U, and V components of the HH coefficient are 0, 2, and 0.
- the run-length number memory Z3021 is initialized to 0, and the encoding target coefficient memory Z3022 is all initialized to 0.
- all the previous line coefficient memories Z3023 are initialized to zero.
- the coefficient extraction coordinates are initialized to (0, 0).
- run length means whether the run length number is non-zero. At (0, 0), since the run length number is zero, run length processing is not performed.
- the flag decoding means reads the variable length code Z403 and decodes the flag information Z4041.
- the significant coefficient decoding means refers to each bit of the flag information memory Z4041, and when a bit rises, considers that the coefficient value corresponding to the bit position is significant, and starts from the code data Z403.
- the variable length code is read, decoded and stored in the coefficient memory Z4053. If no bit is set, zero is written to the coefficient value corresponding to the bit position.
- the run length number decoding means regards the run length number as significant and decodes the run length number. A series of these processes is shown in Z404.
- the run length decoding means decodes the run length from the input code Z403 and stores it in the run length memory Z4051.
- the LH, HL, and HH coefficient values of the same layer 'coordinate (0, 0) in the subband memory Z401 are stored in the coefficient memory Z4053, and the state of each memory becomes Z405.
- the inverse wavelet transform means reads the coefficient value Z406 corresponding to the coordinate (0, 0) from the subband memory Z401, and combines it with the LH, HL, and HH coefficient values of the coefficient memory Z4053 to reverse the two-dimensional H AAR. Perform wavelet transform processing.
- the 0LL component obtained as a result of the inverse wavelet transform is output as pixels A, B, C, D, E, F, G, H, I, J, K, and L in the image memory Z402.
- the LH, HL, and HH coefficient values of the same layer 'coordinate (1, 0) in the subband memory Z401 are stored in the coefficient memory Z4073, and the state of each memory becomes Z407.
- the inverse wavelet transform means reads the coefficient value Z408 corresponding to the coordinate (1, 0) from the subband memory Z401, and combines it with the LH, HL, and HH coefficient values of the coefficient memory Z4073 to obtain a two-dimensional H
- the 0LL component obtained as a result of the inverse wavelet transform is output to the image memory Z402.
- run length means whether the run length number is non-zero. In (4, 0), since the run length is zero, run length processing is not performed.
- the flag decoding means reads the variable-length code Z409 and decodes the flag information Z4101.
- the significant coefficient decoding means and the run length decoding means perform the decoding process as shown in Z410 in the same manner as Z404.
- the value of the flag information memory Z4101 is EAh.
- FLHY second from the top Therefore, the input code Z409 is also decoded to obtain the coefficient value 3, and the coefficient value 3 is stored in the Y component of the LH coefficient in the coefficient memory Z4113.
- FLHU, LHV (third from the highest), FHLU, HLV (fifth from the highest) and FHHU (seventh from the highest) are set.
- the run length decoding means decodes the run length from the input code Z409 and stores it in the run length memory Z4111.
- the LH, HL, and HH coefficient values of the same layer 'coordinate (4, 0) in the subband memory Z401 are stored in the coefficient memory Z4113, and the state of each memory becomes Z411.
- the inverse wavelet transform means reads the coefficient value Z412 corresponding to the coordinate (4, 0) from the subband memory Z401, and combines it with the LH, HL, and HH coefficient values of the coefficient memory Z4113 to obtain a two-dimensional H
- the 0LL component obtained as a result of the inverse wavelet transform is output to the image memory Z402.
- the present invention described above can be applied to uses such as a high-speed and high-quality image distribution service. It can also be applied to applications such as remotely browsing high-quality and high-speed document images stored with a network scanner or FAX multifunction device.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Discrete Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
- Compression Of Band Width Or Redundancy In Fax (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Description
Claims
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US12/309,227 US8463061B2 (en) | 2006-07-13 | 2007-07-13 | Encoding and decoding two-dimensional signal using a wavelet transform |
EP07790733.5A EP2051499B1 (en) | 2006-07-13 | 2007-07-13 | Wavelet-based image encoding and decoding device and encoding method and decoding method |
CN2007800266554A CN101491076B (zh) | 2006-07-13 | 2007-07-13 | 编码和译码装置以及编码方法和译码方法 |
JP2008524848A JP4424522B2 (ja) | 2006-07-13 | 2007-07-13 | 符号化及び復号化装置ならびに符号化方法及び復号方法 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2006192494 | 2006-07-13 | ||
JP2006-192494 | 2006-07-13 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2008007760A1 true WO2008007760A1 (fr) | 2008-01-17 |
Family
ID=38923316
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2007/063952 WO2008007760A1 (fr) | 2006-07-13 | 2007-07-13 | Dispositif de codage et de décodage, procédé de codage et procédé de décodage |
Country Status (5)
Country | Link |
---|---|
US (1) | US8463061B2 (ja) |
EP (2) | EP2051499B1 (ja) |
JP (1) | JP4424522B2 (ja) |
CN (1) | CN101491076B (ja) |
WO (1) | WO2008007760A1 (ja) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2011007478A1 (ja) * | 2009-07-17 | 2011-01-20 | 日本電気株式会社 | ウェーブレット変換符号化/復号方法および装置 |
WO2014013647A1 (ja) * | 2012-07-19 | 2014-01-23 | 日本電気株式会社 | ウェーブレット変換符号化/復号方法および装置 |
US20150156504A1 (en) * | 2009-10-28 | 2015-06-04 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding residual block, and method and apparatus for decoding residual block |
US11539954B2 (en) | 2012-02-29 | 2022-12-27 | Sony Corporation | Image processing device and method with a scalable quantization matrix |
Families Citing this family (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
RU2012117826A (ru) * | 2009-09-30 | 2013-11-10 | Шарп Кабусики Кайся | Устройство увеличения изображения, программа увеличения изображения и устройство отображения |
US8760572B2 (en) * | 2009-11-19 | 2014-06-24 | Siemens Aktiengesellschaft | Method for exploiting structure in sparse domain for magnetic resonance image reconstruction |
RU2012141557A (ru) * | 2010-03-01 | 2014-04-10 | Шарп Кабусики Кайся | Устройство для увеличения изображения, программа для увеличения изображения, запоминающая среда, в которой хранится программа для увеличения изображения, и устройство отображения |
CN102255617B (zh) * | 2010-05-19 | 2014-02-19 | 鸿富锦精密工业(深圳)有限公司 | 哈夫曼树的存储方法及利用数组进行数据解码的方法 |
KR20120035096A (ko) * | 2010-10-04 | 2012-04-13 | 한국전자통신연구원 | 쿼드 트리 변환 구조에서 부가 정보의 시그널링 방법 및 장치 |
TWI595770B (zh) * | 2011-09-29 | 2017-08-11 | 杜比實驗室特許公司 | 具有對稱圖像解析度與品質之圖框相容全解析度立體三維視訊傳達技術 |
JPWO2013065702A1 (ja) * | 2011-11-04 | 2015-04-02 | シャープ株式会社 | 算術復号装置、画像復号装置、算術符号化装置、画像符号化装置、および、算術復号方法 |
CN104662853B (zh) * | 2012-08-21 | 2017-07-11 | 扬智欧洲有限公司 | 软度量值压缩方法 |
CN103152573A (zh) * | 2013-03-15 | 2013-06-12 | 惠州Tcl移动通信有限公司 | 一种移动终端与智能电视间图像帧传输的方法及系统 |
EP3041233A1 (en) * | 2014-12-31 | 2016-07-06 | Thomson Licensing | High frame rate-low frame rate transmission technique |
WO2017046005A1 (en) | 2015-09-18 | 2017-03-23 | Koninklijke Philips N.V. | Method and apparatus for fast and efficient image compression and decompression |
JP7016660B2 (ja) * | 2017-10-05 | 2022-02-07 | キヤノン株式会社 | 符号化装置、その制御方法、および制御プログラム、並びに撮像装置 |
EP3562042A1 (en) * | 2018-04-23 | 2019-10-30 | Total SA | Method of input data compression, associated computer program product, computer system and extraction method |
CN109492648B (zh) * | 2018-09-21 | 2021-12-14 | 云南大学 | 基于离散余弦系数多尺度小波变换的显著性检测方法 |
CN113228109A (zh) * | 2018-12-26 | 2021-08-06 | 松下电器(美国)知识产权公司 | 三维数据编码方法、三维数据解码方法、三维数据编码装置、以及三维数据解码装置 |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH05284368A (ja) * | 1992-03-31 | 1993-10-29 | Fujitsu Ltd | 画像データ符号化・復元方法及びその装置 |
JPH1146372A (ja) * | 1997-05-29 | 1999-02-16 | Sharp Corp | 画像符号化装置及び画像復号装置 |
JP2002112039A (ja) | 2000-09-29 | 2002-04-12 | Ricoh Co Ltd | 画像符号化/復号化処理装置 |
JP2003274185A (ja) | 2002-03-19 | 2003-09-26 | Sanyo Electric Co Ltd | 画像処理方法とその方法を利用可能な画像符号化装置 |
JP2004503964A (ja) * | 2000-06-14 | 2004-02-05 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | カラービデオ符号化及び復号方法 |
JP2005535155A (ja) * | 2001-12-20 | 2005-11-17 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | ビデオシーケンスの圧縮のための符号化方法 |
JP2006187364A (ja) | 2004-12-28 | 2006-07-20 | Daiichi Shokai Co Ltd | 遊技機 |
Family Cites Families (39)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH01146372A (ja) | 1987-12-02 | 1989-06-08 | Sumitomo Electric Ind Ltd | 超電導ショットキゲート電界効果トランジスタ |
JP3158458B2 (ja) * | 1991-01-31 | 2001-04-23 | 日本電気株式会社 | 階層表現された信号の符号化方式 |
US5412429A (en) * | 1993-03-11 | 1995-05-02 | The United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration | Picture data compression coder using subband/transform coding with a Lempel-Ziv-based coder |
US5568598A (en) * | 1994-09-09 | 1996-10-22 | Intel Corporation | Displaying images using progressive fade-in |
US5434567A (en) * | 1994-09-09 | 1995-07-18 | Kabushiki Kaisha Toshiba | Encoding image signals by further transforming transformed DC signals |
US5764802A (en) * | 1994-09-09 | 1998-06-09 | Intel Corporation | Encoding image signals using a blur image |
US5748786A (en) * | 1994-09-21 | 1998-05-05 | Ricoh Company, Ltd. | Apparatus for compression using reversible embedded wavelets |
EP1635579A2 (en) * | 1995-09-12 | 2006-03-15 | Matsushita Electric Industrial Co., Ltd. | Coding method, coding apparatus, wavelet transform apparatus, and wavelet reverse transform apparatus |
US5808683A (en) * | 1995-10-26 | 1998-09-15 | Sony Corporation | Subband image coding and decoding |
US5682152A (en) * | 1996-03-19 | 1997-10-28 | Johnson-Grace Company | Data compression using adaptive bit allocation and hybrid lossless entropy encoding |
US5982434A (en) * | 1996-03-22 | 1999-11-09 | Sony Corporation | Image signal coding method and device thereof, image signal decoding method and device thereof, and recording medium |
US5748116A (en) * | 1996-11-27 | 1998-05-05 | Teralogic, Incorporated | System and method for nested split coding of sparse data sets |
JPH11191153A (ja) * | 1997-12-26 | 1999-07-13 | Ricoh Co Ltd | ウェーブレット変換係数の符号化方法 |
US6208692B1 (en) * | 1997-12-31 | 2001-03-27 | Sarnoff Corporation | Apparatus and method for performing scalable hierarchical motion estimation |
US6148111A (en) * | 1998-04-27 | 2000-11-14 | The United States Of America As Represented By The Secretary Of The Navy | Parallel digital image compression system for exploiting zerotree redundancies in wavelet coefficients |
US6526175B2 (en) * | 1998-06-30 | 2003-02-25 | Sarnoff Corporation | Apparatus and method for packetizing significance-based information |
US6487319B1 (en) * | 1998-11-18 | 2002-11-26 | Sarnoff Corporation | Apparatus and method for identifying the location of a coding unit |
US6356665B1 (en) * | 1998-12-09 | 2002-03-12 | Sharp Laboratories Of America, Inc. | Quad-tree embedded image compression and decompression method and apparatus |
KR100284027B1 (ko) * | 1998-12-22 | 2001-03-02 | 서평원 | 웨이블릿 패킷 계수의 재배치방법 |
JP4367880B2 (ja) * | 1999-12-09 | 2009-11-18 | キヤノン株式会社 | 画像処理装置及びその方法並びに記憶媒体 |
WO2001076255A1 (en) * | 2000-04-04 | 2001-10-11 | Koninklijke Philips Electronics N.V. | Video encoding method using a wavelet transform |
CN1266649C (zh) * | 2000-09-12 | 2006-07-26 | 皇家菲利浦电子有限公司 | 视频编码方法 |
US6898323B2 (en) * | 2001-02-15 | 2005-05-24 | Ricoh Company, Ltd. | Memory usage scheme for performing wavelet processing |
US7409094B2 (en) * | 2001-05-04 | 2008-08-05 | Hewlett-Packard Development Company, L.P. | Methods and systems for packetizing encoded data |
JP4834917B2 (ja) * | 2001-05-14 | 2011-12-14 | 株式会社ニコン | 画像符号化装置および画像サーバ装置 |
JP2003169333A (ja) * | 2001-09-21 | 2003-06-13 | Ricoh Co Ltd | 符号列作成装置、画像伸長システム、画像伸長装置、画像提供システム、符号列作成方法、プログラム及び記録媒体 |
JP2003152544A (ja) * | 2001-11-12 | 2003-05-23 | Sony Corp | データ通信システム、データ送信装置、データ受信装置、および方法、並びにコンピュータ・プログラム |
JP3743384B2 (ja) * | 2002-04-19 | 2006-02-08 | ソニー株式会社 | 画像符号化装置及び方法、並びに画像復号装置及び方法 |
JP4350342B2 (ja) * | 2002-04-26 | 2009-10-21 | 株式会社リコー | 画像処理装置、画像記録装置、カメラシステム、プログラム、記憶媒体及び画像処理方法 |
JP3707456B2 (ja) * | 2002-08-12 | 2005-10-19 | ヤマハ株式会社 | 画像データ圧縮方法および画像データ伸張装置並びに伸張プログラム |
JP3704515B2 (ja) * | 2002-08-30 | 2005-10-12 | Necアクセステクニカ株式会社 | 画像伸張装置及びその方法 |
JP3928859B2 (ja) * | 2002-11-11 | 2007-06-13 | 株式会社リコー | 画像処理装置、画像処理方法、プログラム及び記録媒体 |
US7454069B2 (en) * | 2002-11-20 | 2008-11-18 | Ricoh Company, Ltd. | Image processing apparatus, image compressing apparatus, image decompressing apparatus, image processing method, image compressing method, image decompressing method, information processing apparatus, information processing method, program and recording medium |
US7474792B2 (en) * | 2002-12-02 | 2009-01-06 | Ricoh Company, Ltd. | Image processing apparatus for using distortion amount information |
US7526144B2 (en) * | 2003-02-20 | 2009-04-28 | Ricoh Company, Ltd. | Image processing method, image expansion method, image output method, image conversion method, image processing apparatus, image expansion apparatus, image output apparatus, image conversion apparatus, and computer-readable storage medium |
US7447369B2 (en) * | 2003-03-07 | 2008-11-04 | Ricoh Co., Ltd. | Communication of compressed digital images |
JP4114534B2 (ja) * | 2003-05-02 | 2008-07-09 | ソニー株式会社 | 画像符号化装置及び方法 |
EP1722572B1 (en) * | 2004-02-23 | 2013-02-20 | NEC Corporation | 2-dimensional signal encoding/decoding method and device |
JP2006033507A (ja) * | 2004-07-16 | 2006-02-02 | Sony Corp | 遠隔編集システム、主編集装置、遠隔編集装置、編集方法、編集プログラム、及び記憶媒体 |
-
2007
- 2007-07-13 EP EP07790733.5A patent/EP2051499B1/en not_active Not-in-force
- 2007-07-13 CN CN2007800266554A patent/CN101491076B/zh not_active Expired - Fee Related
- 2007-07-13 US US12/309,227 patent/US8463061B2/en active Active
- 2007-07-13 EP EP15168306.7A patent/EP2942935A1/en not_active Withdrawn
- 2007-07-13 WO PCT/JP2007/063952 patent/WO2008007760A1/ja active Application Filing
- 2007-07-13 JP JP2008524848A patent/JP4424522B2/ja active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH05284368A (ja) * | 1992-03-31 | 1993-10-29 | Fujitsu Ltd | 画像データ符号化・復元方法及びその装置 |
JPH1146372A (ja) * | 1997-05-29 | 1999-02-16 | Sharp Corp | 画像符号化装置及び画像復号装置 |
JP2004503964A (ja) * | 2000-06-14 | 2004-02-05 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | カラービデオ符号化及び復号方法 |
JP2002112039A (ja) | 2000-09-29 | 2002-04-12 | Ricoh Co Ltd | 画像符号化/復号化処理装置 |
JP2005535155A (ja) * | 2001-12-20 | 2005-11-17 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | ビデオシーケンスの圧縮のための符号化方法 |
JP2003274185A (ja) | 2002-03-19 | 2003-09-26 | Sanyo Electric Co Ltd | 画像処理方法とその方法を利用可能な画像符号化装置 |
JP2006187364A (ja) | 2004-12-28 | 2006-07-20 | Daiichi Shokai Co Ltd | 遊技機 |
Non-Patent Citations (2)
Title |
---|
See also references of EP2051499A4 |
SHUITSU MATSUMURA; MICHIAKI KATO; TSUYOSHI TAKEBE: "Wavelet Image Coding Using Power Adaptive Vector Quantization", 1995 IMAGE CODING SYMPOSIUM (PCSJ95), October 1995 (1995-10-01), pages L21 - 122 |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4650592B1 (ja) * | 2009-07-17 | 2011-03-16 | 日本電気株式会社 | ウェーブレット変換符号化/復号方法および装置 |
US8218645B2 (en) | 2009-07-17 | 2012-07-10 | Nec Corporation | Method and apparatus for encoding and decoding by wavelet transform |
WO2011007478A1 (ja) * | 2009-07-17 | 2011-01-20 | 日本電気株式会社 | ウェーブレット変換符号化/復号方法および装置 |
US10136149B2 (en) | 2009-10-28 | 2018-11-20 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding residual block, and method and apparatus for decoding residual block |
US20150156504A1 (en) * | 2009-10-28 | 2015-06-04 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding residual block, and method and apparatus for decoding residual block |
US20150156505A1 (en) * | 2009-10-28 | 2015-06-04 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding residual block, and method and apparatus for decoding residual block |
US10257530B2 (en) | 2009-10-28 | 2019-04-09 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding residual block, and method and apparatus for decoding residual block |
US10178401B2 (en) * | 2009-10-28 | 2019-01-08 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding residual block, and method and apparatus for decoding residual block |
US10154273B2 (en) | 2009-10-28 | 2018-12-11 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding residual block, and method and apparatus for decoding residual block |
US11539954B2 (en) | 2012-02-29 | 2022-12-27 | Sony Corporation | Image processing device and method with a scalable quantization matrix |
WO2014013647A1 (ja) * | 2012-07-19 | 2014-01-23 | 日本電気株式会社 | ウェーブレット変換符号化/復号方法および装置 |
CN104471945B (zh) * | 2012-07-19 | 2018-01-26 | 日本电气株式会社 | 小波变换编码/解码方法和装置 |
JPWO2014013647A1 (ja) * | 2012-07-19 | 2016-06-30 | 日本電気株式会社 | ウェーブレット変換符号化/復号方法および装置 |
US9369734B2 (en) | 2012-07-19 | 2016-06-14 | Nec Corporation | Method and apparatus for encoding and decoding by wavelet transform |
CN104471945A (zh) * | 2012-07-19 | 2015-03-25 | 日本电气株式会社 | 小波变换编码/解码方法和装置 |
Also Published As
Publication number | Publication date |
---|---|
US20090324116A1 (en) | 2009-12-31 |
EP2051499B1 (en) | 2015-09-09 |
JPWO2008007760A1 (ja) | 2009-12-10 |
EP2942935A1 (en) | 2015-11-11 |
EP2051499A1 (en) | 2009-04-22 |
CN101491076A (zh) | 2009-07-22 |
EP2051499A4 (en) | 2012-09-19 |
US8463061B2 (en) | 2013-06-11 |
JP4424522B2 (ja) | 2010-03-03 |
CN101491076B (zh) | 2011-05-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP4424522B2 (ja) | 符号化及び復号化装置ならびに符号化方法及び復号方法 | |
EP0933920B1 (en) | Data processing apparatus and method and storage medium | |
JP3208487B2 (ja) | 情報を圧縮するための装置及び方法 | |
JP2001258027A (ja) | 画像処理装置及び方法及び記憶媒体 | |
JP2004221836A (ja) | 画像処理装置、プログラム、記憶媒体及び符号伸長方法 | |
CA3031546C (en) | Data compression by means of adaptive subsampling | |
JP2005039793A (ja) | 画像符号化装置、プログラム及び記憶媒体 | |
CN108028928B (zh) | 用于快速和高效的图像压缩和解压缩的方法和装置 | |
US7565024B2 (en) | Run length coding and decoding | |
Sandhu et al. | Matlab Based Image Compression Using Various Algorithms | |
CN103561275A (zh) | 一种改进条带波变换的图像压缩方法 | |
KR20020055864A (ko) | 칼라 정지영상의 부호화 및 복호화 방법 | |
JP2004201047A (ja) | 画像処理装置、プログラム及び記憶媒体 | |
Sebastian et al. | Color image compression Using JPEG2000 with adaptive color space transform | |
US20050207659A1 (en) | Image-processing apparatus, program, and recording medium | |
Janaki et al. | Visually improved image compression by using embedded zero-tree wavelet coding | |
JP2006295561A (ja) | 符号化処理装置、符号化処理方法、プログラムおよび記録媒体 | |
Rahman et al. | Efficient Image Compression Technique using JPEG2000 with Adaptive Threshold | |
JP4034236B2 (ja) | 特定変換係数優先符号化装置、特定変換係数優先符号化復号システム及び特定変換係数優先符号化方法 | |
Chebil et al. | Compressed domain editing of JPEG2000 images | |
Makhloufi et al. | QIM watermarking combined to JPEG2000 part I and II. | |
JP2004289737A (ja) | 画像符号化装置、画像符号化プログラム、及び画像復号化プログラム | |
JP4093870B2 (ja) | 画像処理装置、プログラムおよび記憶媒体 | |
Tan et al. | Still image compression based on 2D discrete wavelet transform | |
Bian et al. | Framework of A Hybrid Image Compression Algorithm for View-dependent Visualization. |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200780026655.4 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 07790733 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2008524848 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 12309227 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2007790733 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: RU |