+

WO2001061993A1 - Attenuation de bruit pour signaux video - Google Patents

Attenuation de bruit pour signaux video Download PDF

Info

Publication number
WO2001061993A1
WO2001061993A1 PCT/CA2000/000155 CA0000155W WO0161993A1 WO 2001061993 A1 WO2001061993 A1 WO 2001061993A1 CA 0000155 W CA0000155 W CA 0000155W WO 0161993 A1 WO0161993 A1 WO 0161993A1
Authority
WO
WIPO (PCT)
Prior art keywords
motion
pixels
pixel
still
target
Prior art date
Application number
PCT/CA2000/000155
Other languages
English (en)
Inventor
Rabab K. Ward
Julong Du
Original Assignee
The University Of British Columbia
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by The University Of British Columbia filed Critical The University Of British Columbia
Priority to AU26540/00A priority Critical patent/AU2654000A/en
Priority to PCT/CA2000/000155 priority patent/WO2001061993A1/fr
Publication of WO2001061993A1 publication Critical patent/WO2001061993A1/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N5/00—Details of television systems
    • H04N5/14—Picture signal circuitry for video frequency region
    • H04N5/21—Circuitry for suppressing or minimising disturbance, e.g. moiré or halo
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N5/00—Details of television systems
    • H04N5/14—Picture signal circuitry for video frequency region
    • H04N5/144—Movement detection

Definitions

  • the present invention relates to a noise reducing system for providing a noise reduced value pixels in a video signal.
  • Analog cable television systems are susceptible to additive noise impairments, most commonly, thermal noise, composite second order beats and composite triple beats.
  • Digital cable broadcasting has been introduced to solve these problems.
  • Video noise reduction is one of the proposed solutions.
  • digital image processing related techniques are widely used in reducing video noise.
  • the basic idea underlying most of existing algorithms, and this invention, is to estimate the original value of an impaired signal with redundant data in the neighborhood.
  • correlation may adhere spatially and temporally, in order to maximize the effectiveness of an algorithm, most systems use both spatial and temporal correlated pixels for the calculation.
  • a conventional inter-frame motion detection algorithm assumes a pixel is in a motion area if the difference between the pixel under examination and its corresponding pixel in the history frame is greater than a threshold. It is obvious that the lower the threshold, the less the artifacts are likely to be introduced. However, the lower the threshold, the more chance that a high intensity noise would cause the motion detection logic false alarming.
  • Michael et al. in US patent no. 4,240,106 described two motion detectors which overcome some of the deficiencies of the above described basic motion detector.
  • the Michael et al. detector employs an array of pixel differences disposed about the pixel currently being examined for motion.
  • the respective picture point differences are independently compared to a threshold value to generate bi-level signals corresponding to a majority logic gate which produces a motion signal if the majority of the pixel differences in the array exceed the threshold.
  • the pixel differences of the array of pixels centered about the pixel under examination, are integrated to generate an average difference over the area. The average difference is then compared with a threshold.
  • the object is to differentiate between noise and movement. It is assumed that the larger the number of picture points integrated, the more likely the noise will average to zero whereas movement remains unattended.
  • Roeder et al. in US patent no. 4,661,853 divide the difference matrix formed by target pixel and its surrounding pixels into several sub-matrices, hence improves the property of detail preservation. -However, compared with the Michael et al. detector, Roeder detector is less robust in heavy noise situation as less pixels are involved in the averaging.
  • the present invention relates to a method of reducing noise in a video signal comprising providing a process frame with each of its pixels having a digital value applying a temporal motion detection step in sequence to target pixels in said process frame, said temporal motion detection step being based on difference in values between pixels corresponding to said target pixels in said process frame and a selected number of previous frames to define the status of each said target pixel with respect to the previous frames as a still pixel or a motion pixel, further processing each said motion pixel in an impulse recognition step to further define status of said motion pixels as either still pixels or retain their classification as motion pixels based on status of adjacent pixels in said frames, subjecting said target pixels to a spatial motion detection step based on weighting of statuses of target pixel adjacent pixels in said process frame and at least one of said previous frames to define a motion index for said target pixel.
  • said temporal motion detecting step define status of each said target pixel with respect to each of said previous frame and said spatial motion detection step defines a motion index for said target pixel with respect to each of said previous frames.
  • said method further comprises applying a motion adaptive temporal filter to said still pixels to define a pixel value for each of said still pixel, said motion adaptive temporal filter using said motion index values as thresholds.
  • said method further comprises subjecting said motion pixels to a detailed motion adaptive spatial filter using a dynamic threshold based on said motion index values for said target pixels being processed.
  • Figure 1 is an exemplary block diagram of the non-recursive noise reduction system including temporal motion adaptive and spatial motion/detail adaptive processing circuits
  • Figures 2, 3, 4 describe the motion detection, still correction and motion post processing circuits.
  • Figure 5 is a logic schematic diagram of exemplary circuits for the adaptive temporal moving average filter implementation.
  • Figure 6 is the pictorial representations of arrays of points presenting the pixels involved in processing and the mathematical expression of the processing algorithm.
  • Figure 7 is an exemplary circuit and a graphic representation that shows a typical relationship between the motion probability and the threshold value for the spatial soft switching.
  • Figure 8 is an exemplary logic schematic diagram of the parallel processing circuit. It is for 5x3 array processing, it can be easily expanded to mxn array processing.
  • Figure 9 demonstrates a digital TV set-top decoder application specific circuit design including DRAM paging and address generation block diagrams. This circuit can replace the frame delays and the field delay in Figure 1.
  • Figure 10 presents application examples that propose how the processing unit fits into a set-top box with analog and digital demodulation, respectively.
  • Figure 11 is a schematic flow diagram illustrating the method of the present invention.
  • the system does noise reduction pixel by pixel.
  • the pixel being processed is called the target pixel
  • the pixels which have the same position in the history frames are called the corresponding pixels of the target pixel in those frames.
  • the system also utilizes data from previous frames, in the following description of an example of the invention, three frames which are called frame 1, frame 2 and frame 3 in this document, where frame 1 is the nearest past frame, and frame 3 denotes the furthest past frame.
  • Figure 3A illustrates the correction logic which will be described in more detail later.
  • the best feature of this scheme is that it will detect very fine detail motion, although it is based on an assumption similar to that of previous inventions that motion will not likely take place in an isolated pixel.
  • the above "still” detection circuit is unfortunately sensitive to motion in areas with fine details and to ringing effects of moving sharp edges. Ringing effects appear only in vertical direction, however, motion-still transition area tends to create after image in both vertical and horizontal directions.
  • a post-processor is placed after the temporal motion detection to protect moving boundaries. This processing is also effective on a moving object that has a gradual change in its gray level intensity. Different from an average filter, this circuit integrates the product of motion status value and the reciprocal of its distance from the target pixel to estimate the motion index of the pixel (motion index is an indicator of the probability of a pixel being a motion pixel).
  • multiplication operations are involved, as they are one bit operations, it is not as complicated as it appears.
  • An edge/detail adaptive spatial filter follows the above temporal filter.
  • An edge/detail detection circuit examines the pixels in a spatial window with the target pixel in the center and marks the pixels that have a value close enough to that of the center target pixel. The filter averages the marked pixel values and replaces the value of the target pixel by this average. Based on the finding that human eyes are less sensitive to noise in moving details in a picture and that the lower the threshold, the less the chance of possible blurring but the less the noise reduction, the threshold for deciding if the value of a pixel in the window is close enough to that of the target pixel is made motion adaptive, i.e. a higher-than-necessary threshold for motion areas and a lower threshold for still areas are chosen.
  • the spatial filtering tends to compensate for the difference in noise reduction between the motion and the non-motion areas resulting from our previous temporal still-only filtering.
  • Soft switching between the motion threshold and the still threshold is employed here to make the image after processing more natural-looking.
  • the first step as indicated at 10 is to grab a frame and convert it from analog to digital, if necessary, and provide the digital output with a gray scale value for each of the pixels in the frame.
  • the information on the process frame i.e. the frame currently being processed, is fed as indicated by the arrow 12 to a temporal motion detection step 14 to which is also fed from memory 16 as indicated by arrow 18 data on previous frames which for the purpose of this description, designated as frames 1, 2 and 3.
  • frames 1, 2 and 3 any reasonable number of such previous frames may be used in the process.
  • the temporal motion detection step 14 and its operation will be described in more detail hereinbelow with respect to Figure 2.
  • This step in the process is intended to provide a first definition of the pixel status of the target pixel being processed and outputs the pixel as being either still as indicated at 20 or positive motion at 22 or negative motion as indicated at 24. Obviously, mistakes may be made and to better ensure that those pixels as indicated at 22 and 20 as containing motion are further processed in an impulse spatial pattern recognition step as indicated at 28 (as described more fully with respect to the example shown in Figure 3) to again further define the status of the target pixel as still as indicated at 30 or motion as indicated at 32.
  • the pixels are further process to ensure proper values are applied to the pixels before they are reconverted to analog for video projection. This is accomplished in the present invention by processing the pixels whose status has been defined as indicated at 20, 30 and 32 in the spatial motion detection system 34 to define a motion index value as indicated at 36 based on weighted averages for adjacent pixels to the target pixels in the process frame and proceeding frames as described in more detail in the following example with respect to Figure 4.
  • the motion index value 36 is then used in conjunction with the motion adaptive temporal filter as indicated at 40 to process the pixels and provide a pixel value 42 for each of the target pixels as will be described in more detail hereinbelow with respect to Figure 5.
  • the motion index value 36 is used to set a dynamic threshold 44 which is applied to the pixel values 42 in a detailed motion adaptive spatial filter system or step 46 to define a final pixel value 48 for each of the target pixels which is then use to produce the picture.
  • the detailed motion adaptive special filter 46 is described in more detail in conjunction with Figures 6, 7 and 8.
  • the present invention comprises a motion adaptive temporal filter
  • Figure 5 detail/motion adaptive spatial filter
  • Figures 6 and 7 detail/motion adaptive spatial filter
  • Figures 2, 3, 4 a motion detector with impulse pattern recognition and motion boundary protection circuitry
  • Figure 10 shows two examples that fit the noise reduction device into digital TV set-top decoders with digital and analog demodulation circuits, respectively.
  • Figure 1 is a block diagram showing the system structure of this invention for video noise reduction.
  • reference numeral 101 denotes a digitized component video signal as the input to the noise reduction system.
  • the three cascaded frame buffers delay the input so that at anytime signals from the past three frames are also available.
  • Reference numeral 111, 121 and 131 denote the signal of framel, frame2 and frame3, respectively.
  • the four signals 101, 111, 121 and 131 are supplied to the motion detection circuit and the motion adaptive temporal filter.
  • the motion detection circuit consists of three function blocks: temporal motion detection, impulse pattern recognition and spatial motion detection. Each block accomplishes one stage of motion detection and is described in detail in Figures 2, 3 and 4, respectively.
  • the motion detection circuit outputs signal 106 which controls the motion adaptive temporal filter and outputs signal 104 which is further converted to the threshold 109 used in the detail/motion adaptive spatial filter ( Figures 6, 7).
  • the output from the temporal filter ( Figure 5) is fed into a delay structure which delays the incoming signal to five vertically successive lines in a frame.
  • Signals 108 1, 108_3 and 108_5 are in the current field
  • signals 108_2 and 108_4 are in the field which is delayed by one field from the current field. Since the TV signal is interlaced, in a frame, the line where signal 108_2 abides appears between the two lines where signal 108 1 and signal 108_3 abide respectively.
  • the output from the spatial filter, signal 107 is the final processed output signal of this invented system. Motion Detection
  • the motion detection circuit is described more specifically as follows, referring to Figures 2, 3 and 4.
  • the detection consists of three stages: temporal motion detection, impulse pattern recognition and spatial motion detection.
  • Temporal motion detect circuit detects motion by comparing the differences between the target pixel and its corresponding pixel in each of the past three frames with a threshold.
  • the subtracter circuit 200 subtracts the target pixel signal 101 from signals 111, 121 and 131, which are its corresponding pixels in framel, frame2 and frame3, respectively.
  • the difference signals 201, 202 and 203 are applied to the comparator circuit 210, where they are compared with a threshold 103.
  • the input- output relationship of each comparator circuits can be described by the equation below, where a difference signal 201 or 202 or 203 is denoted as ⁇ . still if
  • Motion Mode positive motion if ⁇ > + threshold negative motion if ⁇ ⁇ - threshold
  • Motion mode may take one of 3 values, motion status as will be more full discussed below may take on 2 values.
  • the motion mode for each pixel could be one of the following three: still, positive motion and negative motion.
  • the outputs 211, 212 or 213 from the comparator circuits 210 are 2 ⁇ bit signals with one bit indicating whether or not the target pixel is in motion, and the other bit indicates the polarity of the motion if the target pixel is in motion.
  • the temporal motion detection ( Figure 2) is sensitive to noise, which means noise in the video signal may cause errors in the motion detection results. With noise superimposed, a still pixel can be falsely detected as a motion pixel, and on the other hand, a motion pixel can also be falsely detected a still pixel.
  • the function of the impulse pattern recognition logic ( Figure 3) is to correct the error when a still pixel is mistaken as a motion pixel.
  • the theory that the impulse pattern recognition is based on is that higher-than-threshold noise appears randomly in the image as impulses, but video images are coherent which means each pixel is related to its surrounding pixels. By examining the correlation of the target pixel with its surrounding pixels, the wrongly detected motion caused by noise can be fixed.
  • the pattern recognition circuit analyzes the neighboring area of the target pixel e.g. a 5x3 window, forms a 5x3 matrix of the motion modes of each pixel and detects if the matrix fits one of the six patterns shown in Figure 3.
  • the central element in each of these matrixes corresponds to the target pixel or its corresponding pixel.
  • the six patterns represent the cases when the motion status of one pixel or two adjacent pixels in the center of the window is not coherent with the motion status of their surrounding pixels.
  • the central pixel is in positive motion, and their surrounding pixels are either still or in negative motion, or, when the central pixel is in negative motion, and their surrounding pixels are either still or in positive motion
  • the motion of the central pixel is considered to be erroneous i.e. the motion of the target pixel is considered to have been falsely detected by the temporal motion detection circuit. It is most likely caused by high peak noise.
  • a 5x3 matrix whose central pixel is the target pixel can be implemented e.g. by using pixel delay elements 360s and line delay elements 350s, which are constructed as in Figure 3.
  • the matrix elements 300 to 314 are the values of the motion modes with respect to framel of the pixels in the 5x3 window of the present frame. They are supplied to a combinational logic 370.
  • the logic 370 performs the pattern recognition with the six different patterns, and corrects the false motion detection, if there is any.
  • the 1— bit output signal 371 indicates the decision on the motion status of the target pixel: either a still status or a motion status.
  • the spatial motion detection circuit calculates a motion index value on an mxn matrix whose elements are equal to the values 0 or 1 and represent the motion status of the pixels.
  • the central element in this matrix is the motion status of the target pixel.
  • Figure 4 shows an example of spatial motion detection on a 5x3 matrix.
  • the 5x3 matrices can be implemented e.g. using pixel delay elements 460s and line delay elements 450s, which are constructed as in Figure 4.
  • the weight factors can be properly adjusted for better performance according to the specific application.
  • the weight factor for a pixel is inversely proportional to the distance between this pixel and the target pixel.
  • the weighted elements are added together in adder 480, and the sum 106_1 forms the "motion index" value of the target pixel with respect to framel.
  • Motion indices values for the target pixel with respect to frame2 and frame3, denoted as 106_2 and 106_3, are obtained by similar circuits.
  • 106 Along with 106 1, they are grouped and denoted as 106 in Figure 1, and are used to control the motion adaptive temporal filter.
  • the average of 106_1, 106_2 and 106_3 which is denoted as 104 is used to control the detail/motion adaptive spatial filter.
  • the motion index value for a pixel does not only depend on its own motion status, it also depends on the motion status of its surrounding pixels. Thus if a pixel is detected as a still pixel while its surrounding pixels are in motion, then its motion index value will be high indicating it has higher probability of being a motion pixel and less probability of being a still pixel. This case often happens when noise is present. Therefore, the spatial motion detection represents a more accurate motion estimation mechanism which is especially effective in protecting the boundary of a moving area from artifacts.
  • noise reduction of this filter is based on the facts that, in stationary video sequences, noise causes the only difference between successive frames, and the noise is random with zero mean.
  • the temporal filter averages the target pixel with those corresponding pixels in the previous frames for which it is relatively detected as a still pixel, and therefore achieves noise reduction without impairing the spatial resolution of the picture.
  • Each motion index value 106 1, 106_2 and 106_3 from the motion detection circuit is compared with a threshold 102.
  • the comparison results are denoted as signal 501, 502 and 503 for framel, frame2 and frame3, respectively.
  • the signal 101 from present frame is directly connected to the adder 540 as one of its inputs.
  • Video signals 111, 121 and 131 from the past three frames are connected to the adder circuit 540 via three switches 521, 522 and 523, respectively. When a switch is on, it allows the video signal to pass to the adder.
  • the switches are controlled by the combinational logic of the three motion index values 511, 512 and 513.
  • the logic functions are as follows.
  • the comparison result 501, 502 or 503 is set to high when its corresponding motion index value is greater than the threshold 102; otherwise, it is set to low.
  • the switch 521 is on only if 511 is low; otherwise, it is off.
  • the switch 522 is on only if both 511 and 512 are low; otherwise, it is off.
  • the switch 523 is on only if 511, 512 and 513 are low, otherwise, it is off.
  • the sum 541 of the adder is supplied to the divider 550, and the divisor is the number of signals included in the adder 540.
  • the quotient 105 of the divider is the result of the motion adaptive temporal filter, and this output would be the final filtered value of the target pixel or it would be fed into the detail/motion adaptive spatial filter. Detail/Motion Adaptive Spatial Filter
  • the algorithm for the detail/motion adaptive spatial filter is illustrated in detail in Figures 6 and 7.
  • the spatial filter operates on a mxn image window, where the m and n denotes the number of rows and the number of columns of the window respectively, with the target pixel as the central pixel.
  • Figure 6A illustrates a 5x5 window with the pixel at the location of m-th row and the n-th column as the target pixel. The circles represent pixels in the 5x5 window and the black one is the target pixel. Each pixel in the window is compared with the target pixel first.
  • the spatial filter adds up all the pixels whose absolute difference with the target pixel is smaller than a threshold, and computes the average of all those pixels.
  • FIG. 6B shows a mathematical expression of the algorithm discussed above.
  • Aggressive spatial filtering (high value for the threshold) may cause smoothing effects, which appear as blurring of the picture.
  • the level of the threshold is a trade off between preserving the fine details and reducing noise.
  • a common practice is to sacrifice noise reduction by setting the threshold to a fairly low level.
  • the preferred circuit varies the threshold level as a function of the motion index value. The relationship between the threshold level and the motion index value is shown in Figure 7.
  • the motion index value is expected to be low, and vice versa.
  • the threshold is set to a low level, therefore, the spatial filter does little noise reduction on still parts of the picture, and tries to preserve all the fine details.
  • a small threshold results in less noise reduction, the overall noise reduction on these still parts of image is maintained because the temporal filter has achieved its maximum noise reduction on those pixels.
  • more aggressive spatial noise reduction process is done by increasing the threshold to high values.
  • the motion index value is between the high value and the low value, i.e. the target pixel is in an the area between motion and still the threshold is adjusted in between linears.
  • FIG. 7 illustrates an exemplary circuitry which implements the attributes of the detail/motion adaptive spatial filter of Figure 6.
  • This example circuit operates on a five line by three column sub-array of the image filtered by the temporal filter of Figure 5.
  • Data inputs labeled 800_1 to 800_5 correspond to the line delayed signals 108_ 1 to 108_5 of Figurel. They represent vertically aligned five adjacent pixels in a frame.
  • Both 800_3 and 800_3D, (suffix D indicates the signal is one-sample-period delayed), are considered as reference pixels.
  • Fourteen data selectors, CI .. C14 (each consists of a subtracter 810, a comparator 820 and a data switch 830) are all similar and operate as follows. Refer to one of the data selector CI, the value of one of the reference pixels is coupled to the positive terminal of the subtracter 810 while one of the non-reference pixels is applied to the negative terminal of the subtracter 810. The difference produced by the subtracter 810 is then coupled to comparator 820 where it is compared with threshold 109. Comparator 820 generates a logic one output if the absolute value of the difference is smaller than the threshold 109. The logic one will control switch 830 to let the data pass through for summation at the adder 840. If the absolute value of the difference is greater than the threshold 109, comparator 820 will output a zero to the control terminal of the switch 830 and turn off the switch. The data selector will then output a zero instead of passing the data through.
  • 800_3 is smaller than the threshold 109 among the five data 800_1D to 800_5D and four data 800_1, 800_2, 800_4, 800_5 to sum together at the adder 850.
  • Five data selectors choose data that the difference with reference pixel 800__3D is less than the threshold 109 among the five data 800_1 to 800_5D to sum together at adder 840.
  • Result from adder 850 is delayed by one sample period and then add with result from adder 840. This is the sum of all pixels in a five line by three columns sub-array window with reference pixel 800_3 in the center, which has a value close enough to the value of the center pixel 800_3. This sum is further divided by the number of pixels that have been summed up and obtain the average value 107.
  • Signal 107 is the output of the invented system.
  • FIG 9 shows an example of the implementation of our noise reduction system using dynamic random access memory (DRAM) or synchronous dynamic access memory (SDRAM), which is a popular choice for today's set-top converter design.
  • DRAM dynamic random access memory
  • SDRAM synchronous dynamic access memory
  • Signals 101, 111, 121, 131, 108_4 and 105 are corresponding to the signals of the same name in Figure 1.
  • this particular example circuit operates at four times of the speed of the pixel rate and divide the SDRAM DRAM in to four segments for the three frames and one field storage.
  • Figure 10A shows the interface with an set-top digital decoder with analog demodulation.
  • Figure 10B is an example interface with set-top video decoder with digital demodulation.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Picture Signal Circuits (AREA)

Abstract

L'invention concerne un procédé d'atténuation de bruit pour signaux vidéo qui consiste à traiter les pixels d'une trame de traitement d'un signal vidéo en conjonction avec un nombre choisi de trames traitées préalablement pour déterminer les valeurs pour les pixels cibles dans la trame de traitement, et ce par l'application d'un détecteur de mouvement temporel pour déterminer si un pixel cible a un statut de repos ou de mouvement et pour continuer le traitement des pixels ayant un statut de mouvement au stade de reconnaissance des impulsions afin de déterminer ultérieurement si le pixel doit être classé comme possédant un statut de repos ou de mouvement. Les pixels sont ensuite vérifiés à un stade de détection de mouvement spatial, et ce pour obtenir une valeur d'index de mouvement pour les pixels; cette valeur d'index de mouvement est utilisée dans un filtre adaptatif temporel de mouvement pour définir la valeur de pixel du pixel de repos. La valeur d'index de mouvement des pixels est utilisée pour fixer un seuil dynamique dans un filtre spatial actif de mouvement détaillé afin de définir la valeur de pixels pour les pixels de mouvement. Le système peut être intégré à un boîtier de convertisseur pour convertir les signaux numériques en signaux analogiques à des fins d'utilisation conjointe avec les services de télévision analogique utilisant des signaux numériques.
PCT/CA2000/000155 2000-02-17 2000-02-17 Attenuation de bruit pour signaux video WO2001061993A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
AU26540/00A AU2654000A (en) 2000-02-17 2000-02-17 Noise reduction for video signals
PCT/CA2000/000155 WO2001061993A1 (fr) 2000-02-17 2000-02-17 Attenuation de bruit pour signaux video

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CA2000/000155 WO2001061993A1 (fr) 2000-02-17 2000-02-17 Attenuation de bruit pour signaux video

Publications (1)

Publication Number Publication Date
WO2001061993A1 true WO2001061993A1 (fr) 2001-08-23

Family

ID=4143043

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CA2000/000155 WO2001061993A1 (fr) 2000-02-17 2000-02-17 Attenuation de bruit pour signaux video

Country Status (2)

Country Link
AU (1) AU2654000A (fr)
WO (1) WO2001061993A1 (fr)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1383311A1 (fr) * 2002-07-16 2004-01-21 Broadcom Corporation Techniques numériques de réduction du bruit
EP1633130A1 (fr) * 2003-06-10 2006-03-08 Sony Corporation Recepteur de television et procede de traitement d'images
US8184705B2 (en) 2008-06-25 2012-05-22 Aptina Imaging Corporation Method and apparatus for motion compensated filtering of video signals
US9667964B2 (en) 2011-09-29 2017-05-30 Dolby Laboratories Licensing Corporation Reduced complexity motion compensated temporal processing
US10971192B2 (en) 2018-11-07 2021-04-06 Genetec Inc. Methods and systems for detection of anomalous motion in a video stream and for creating a video summary
CN113438488A (zh) * 2021-06-18 2021-09-24 北京锐马视讯科技有限公司 低码率视频优化编码方法和装置、设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0455434A2 (fr) * 1990-04-28 1991-11-06 Sony Corporation Réducteurs de bruit
EP0502615A2 (fr) * 1991-03-07 1992-09-09 Matsushita Electric Industrial Co., Ltd. Procédé de détection de mouvement dans un signal vidéo et réducteur de bruit avec utilisation de ce procédé
EP0578311A2 (fr) * 1992-07-10 1994-01-12 Koninklijke Philips Electronics N.V. Appareil d'examen à rayons X et moyens de diminution de bruit destinés à être utilisés dans un tel appareil
EP0643533A2 (fr) * 1993-06-07 1995-03-15 Texas Instruments Incorporated Circuit détecteur de mouvement d'image et circuit de réduction de bruit
WO1997039572A1 (fr) * 1996-04-12 1997-10-23 Snell & Wilcox Limited Reducteur de bruit video

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0455434A2 (fr) * 1990-04-28 1991-11-06 Sony Corporation Réducteurs de bruit
EP0502615A2 (fr) * 1991-03-07 1992-09-09 Matsushita Electric Industrial Co., Ltd. Procédé de détection de mouvement dans un signal vidéo et réducteur de bruit avec utilisation de ce procédé
EP0578311A2 (fr) * 1992-07-10 1994-01-12 Koninklijke Philips Electronics N.V. Appareil d'examen à rayons X et moyens de diminution de bruit destinés à être utilisés dans un tel appareil
EP0643533A2 (fr) * 1993-06-07 1995-03-15 Texas Instruments Incorporated Circuit détecteur de mouvement d'image et circuit de réduction de bruit
WO1997039572A1 (fr) * 1996-04-12 1997-10-23 Snell & Wilcox Limited Reducteur de bruit video

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1383311A1 (fr) * 2002-07-16 2004-01-21 Broadcom Corporation Techniques numériques de réduction du bruit
US7050501B2 (en) 2002-07-16 2006-05-23 Broadcom Corporation Digital noise reduction techniques
EP1633130A1 (fr) * 2003-06-10 2006-03-08 Sony Corporation Recepteur de television et procede de traitement d'images
EP1633130A4 (fr) * 2003-06-10 2006-06-14 Sony Corp Recepteur de television et procede de traitement d'images
US7953282B2 (en) 2003-06-10 2011-05-31 Sony Corporation Television receiver and picture processing method
US8184705B2 (en) 2008-06-25 2012-05-22 Aptina Imaging Corporation Method and apparatus for motion compensated filtering of video signals
US9667964B2 (en) 2011-09-29 2017-05-30 Dolby Laboratories Licensing Corporation Reduced complexity motion compensated temporal processing
US10971192B2 (en) 2018-11-07 2021-04-06 Genetec Inc. Methods and systems for detection of anomalous motion in a video stream and for creating a video summary
US11893796B2 (en) 2018-11-07 2024-02-06 Genetec Inc. Methods and systems for detection of anomalous motion in a video stream and for creating a video summary
CN113438488A (zh) * 2021-06-18 2021-09-24 北京锐马视讯科技有限公司 低码率视频优化编码方法和装置、设备及存储介质

Also Published As

Publication number Publication date
AU2654000A (en) 2001-08-27

Similar Documents

Publication Publication Date Title
US6061100A (en) Noise reduction for video signals
EP0951693B1 (fr) Interpolation diagonale adaptative permettant d'ameliorer la resolution d'une image
US4691230A (en) Motion vector detection system of a moving object on a screen
US6262773B1 (en) System for conversion of interlaced video to progressive video using edge correlation
US6681058B1 (en) Method and apparatus for estimating feature values in a region of a sequence of images
CN101309385B (zh) 一种基于运动检测的去隔行处理方法
US8189105B2 (en) Systems and methods of motion and edge adaptive processing including motion compensation features
WO2014114098A1 (fr) Procédé et dispositif d'évaluation de qualité vidéo dans le domaine temporel côté terminal
EP0878776B1 (fr) Réduction adaptative intra-trame du bruit gaussien utilisant la logique floue
KR20060076176A (ko) 비디오 시퀀스에서 시간적 잡음 제거 방법 및 시스템
JP4213035B2 (ja) オクルージョン領域を検出するオクルージョン検出器および方法
JP3291597B2 (ja) 動き検出回路及びノイズ低減回路
US5189710A (en) Method and an apparatus for generating a video binary signal for a video image having a matrix of pixels
JP2003224854A (ja) 動きベクトル検出装置及び画像処理装置並びにコンピュータ・ソフトウエア
US20100079665A1 (en) Frame Interpolation Device
JPH0226484A (ja) テレビ画像の動き検出装置
JP2826018B2 (ja) ビデオ信号のノイズ低減システム
WO2001061993A1 (fr) Attenuation de bruit pour signaux video
JPH0981754A (ja) 動き検出回路
US20200265557A1 (en) Motion detection method and image processing device for motion detection
JP2839536B2 (ja) 動き検出方法および動き検出器
KR100351160B1 (ko) 영상 움직임 보상 장치 및 방법
US20100158320A1 (en) Method of motion detection using content adaptive penalty function
JPH06319082A (ja) 傷検出回路および傷補正回路
KR100949137B1 (ko) 영상 보간 장치와 그 방법 및 그 방법이 기록된 컴퓨터로읽을 수 있는 기록매체

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AL AM AT AU AZ BA BB BG BR BY CA CH CN CR CU CZ DE DK DM EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): GH GM KE LS MW SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

122 Ep: pct application non-entry in european phase
点击 这是indexloc提供的php浏览器服务，不要输入任何密码和下载