+

CN108898611B - An Active Contour Segmentation Model for Blurred Regions Based on Salient Perceptual Priors - Google Patents

An Active Contour Segmentation Model for Blurred Regions Based on Salient Perceptual Priors Download PDF

Info

Publication number
CN108898611B
CN108898611B CN201810630459.8A CN201810630459A CN108898611B CN 108898611 B CN108898611 B CN 108898611B CN 201810630459 A CN201810630459 A CN 201810630459A CN 108898611 B CN108898611 B CN 108898611B
Authority
CN
China
Prior art keywords
image
fuzzy
follows
value
salient
Prior art date
Legal status (The legal status 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 status listed.)
Active
Application number
CN201810630459.8A
Other languages
Chinese (zh)
Other versions
CN108898611A (en
Inventor
方江雄
刘花香
柳和生
顾华奇
刘军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
East China Institute of Technology
Original Assignee
East China Institute of Technology
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 East China Institute of Technology filed Critical East China Institute of Technology
Priority to CN201810630459.8A priority Critical patent/CN108898611B/en
Publication of CN108898611A publication Critical patent/CN108898611A/en
Application granted granted Critical
Publication of CN108898611B publication Critical patent/CN108898611B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/149—Segmentation; Edge detection involving deformable models, e.g. active contour models
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/12—Edge-based segmentation

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Image Analysis (AREA)

Abstract

一种基于显著感知先验的模糊区域型活动轮廓分割模型，主要包括水平集函数的定义、能量泛函的构建和基于能量泛函的求解过程。通过将图像局部区域信息和显著性先验信息融合于模糊主动轮廓模型，构建了区域型模糊项和显著感知先验模糊项的凸能量泛函，并非用欧拉‑拉格朗日公式而是直接计算能量的变化值来更新水平集函数，不仅提高了分割灰度不均匀图像和弱边缘图像的分割效果，而且还使得分割结果与初始条件无关。

Figure 201810630459

A fuzzy region-based active contour segmentation model based on saliency perception prior mainly includes the definition of level set function, the construction of energy functional and the solution process based on energy functional. By fusing the image local area information and saliency prior information into the fuzzy active contour model, the convex energy functional of the regional fuzzy term and the saliency perceptual prior fuzzy term is constructed. Instead of using the Euler-Lagrangian formula, Directly calculating the change value of energy to update the level set function not only improves the segmentation effect of the image with uneven grayscale and weak edge, but also makes the segmentation result independent of the initial conditions.

Figure 201810630459

Description

Fuzzy region active contour segmentation model based on significant perception prior
Technical Field
The invention relates to a method for image segmentation in the technical field of image processing, in particular to a Fuzzy Active Contour with Saliency-aware boundary segmentation model based on significant perception Prior.
Technical Field
Image segmentation is one of the most important tasks in the field of image processing and computer vision, and the purpose of the image segmentation is to extract region objects with the same characteristics in an image. An image segmentation model based on Active Contour (Active Contour) has become a research hotspot in the field of image segmentation in recent years, because the initial estimation state of the model and the prior knowledge of image data can be unified in the feature extraction process, and the prior knowledge obtained in the segmentation process can be used for guiding the segmentation process. However, the classical Chan-Vese segmentation model requires periodic re-initialization of the level set function during curve evolution, thereby increasing the amount of computation and numerical computation error. Based on a Fuzzy Energy-based Active Contour model, Fuzzy Energy is introduced into an Active Contour, a traditional Euler-Lagrange equation is not adopted, but a fast optimization algorithm is adopted to directly minimize a Fuzzy Energy function, convergence can be achieved in a limited iteration process, and a reinitialization process is avoided. Therefore, the active contour model based on the fuzzy energy becomes one of the research hotspots in the image segmentation field in recent years.
It was found through a search of the prior art literature that the local binary fitting-based regional level set method segmented the grayscale inhomogeneous Image (Li c.m., Kao c.y., Gore j.c., and Ding z. "Minimization of region-scalable positioning for Image segmentation", IEEE Transition on Image Processing (2008)17: 1940-. A Saliency-driven region edge top-down level set segmentation model (Zhi X.H., Shen H.B.) "Alimentary drive region-edge-based bottom level set evolution of Saliency-driven region edge reveals asynchronous focus in image segmentation, SDREL model for short), Pattern registration (2018)80:241-255), an energy function is established by fusing a Saliency map and color gray, then a rule item is constructed for edge extraction, but the segmentation of complex gray non-uniform images still has the problem of noise interference. However, these image segmentation models are not locally optimal solutions to the convex energy functional, making the segmentation result dependent on the initialization conditions, and more importantly, the weak edge images cannot be accurately segmented.
Disclosure of Invention
The invention aims to mainly solve the problem that the existing segmentation model is difficult to accurately segment gray-scale uneven images and weak boundary images by providing a fuzzy regional active contour based on significant perception prior. In addition, the convex optimization energy functional of the model enables the segmentation result to be independent of the initial condition, and the convergence can be achieved only for limited times.
The technical scheme of the invention is as follows: the method has the advantages that the local area information and the significance prior information of the image are fused with the fuzzy active contour model, the convex energy functional of the area type fuzzy term and the significance perception prior fuzzy term is constructed, the Euler-Lagrangian formula is not used, the change value of energy is directly calculated to update the level set function, the segmentation effect of segmenting the image with uneven gray level and the image with weak edge is improved, and the segmentation result is independent of the initial condition. The method comprises the following specific steps:
step 1: the level set function u defines:
Figure GDA0003160078700000021
wherein, I is an image domain, x is a pixel point, and C is a closed curve (a pseudo-zero level set function) in the image domain omega.
Step 2: and (5) constructing an energy functional. The energy functional is composed of a regional fuzzy term and a significant perception prior term, wherein the regional fuzzy term EimgThe definition is as follows:
Eimg(u)＝λ∫Ω[u(x)]m(S(x)-c1)2dx+λ∫Ω[1-u(x)]m(S(x)-c2)2dx
where λ is a weight coefficient greater than 0, c1And c2Is the image pixel mean, u (x) is the fuzzy membership function, m is the weight power exponent (which can be 1 or 2), and s (x) is the local area map. For simplicity of calculation, S (x) is defined as a scale r1Variance σ1The image is a Gaussian convolution characteristic map
Figure GDA0003160078700000025
Wherein
Figure GDA0003160078700000026
Is of the scale r1Variance σ1The image gaussian function of (2). Image grey scale value c1And c2The definition is as follows:
Figure GDA0003160078700000022
the significant perceptual prior term is defined as follows:
Esal(u)＝α∫Ωg(h(x)-s1)2·[u(x)]mdx+α∫Ωg(h(x)-s2)2·[1-u(x)]mdx
where α is a weight coefficient greater than 0, s1And s2Is the salient feature image pixel mean, h (x) is the salient detection function, which is defined as
Figure GDA0003160078700000023
IuIs the average of the regional features of the image,
Figure GDA0003160078700000024
is of the scale r2Variance σ2G is an edge detection operator, which is defined as follows:
Figure GDA0003160078700000031
where g is a square matrix of odd rows N,
Figure GDA0003160078700000032
k is the number of matrix elements as the distance from the current point to the neighboring point. Salient feature image pixel mean s1And s2Is defined as:
Figure GDA0003160078700000033
therefore, the energy function expression of the segmentation model is as follows:
E(u)＝λ∫Ω[u(x)]m(S(x)-c1)2dx+λ∫Ω[1-u(x)]m(S(x)-c2)2dx
+α∫Ωg(h(x)-s1)2·[u(x)]mdx+α∫Ωg(h(x)-s2)2·[1-u(x)]mdx
and step 3: the fuzzy membership function u (x) in the iterative process is calculated. For calculating the level set function u (x), the parameter c is fixed1、c2、s1And s2And calculating the minimum value of the energy functional, and obtaining the minimum value through an Euler-Lagrange formula:
λ∫Ω[u(x)]m(S(x)-c1)2dx+λ∫Ω[1-u(x)]m(S(x)-c2)2dx
+α∫Ωg(h(x)-s1)2·[u(x)]mdx+α∫Ωg(h(x)-s2)2·[1-u(x)]mdx＝0
from the above formula, the variable u (x) can be derived as follows:
Figure GDA0003160078700000034
and 4, step 4: and (5) solving an energy functional. According to Fuzzy energy-based active constraint proposed by Krinidis and Chatzis on IEEE Transaction on Image Processing and a computing method proposed by Song and Chan in A fast algorithm for level set based optimization proposed on report CAM-UCLA, the energy functional of the segmentation model is directly computed, and the value of membership degree u (x) is updated by judging whether the energy functional of all pixel points in the Image domain changes. To calculate the difference Δ E of the energy functional, first the constant c is calculated1、c2、s1And s2The update value of (2). Suppose P is a certain pixel point in the image, and the corresponding gray value is s0And degree of membership of u0. Correspondingly, the new degree of membership to the same fixed point P is unMean value of image gray scale c1And c2Become two new values
Figure GDA0003160078700000035
And
Figure GDA0003160078700000036
the calculation process is as follows:
Figure GDA0003160078700000041
wherein t is1＝∑Ω[u(x)]m。
Thus, the new gray scale value of the pixel point P in the image
Figure GDA0003160078700000042
Is composed of
Figure GDA0003160078700000043
Similarly, calculating new pixel value of pixel point P in image
Figure GDA0003160078700000044
New pixel values in saliency map
Figure GDA0003160078700000045
And
Figure GDA0003160078700000046
respectively as follows:
Figure GDA0003160078700000047
Figure GDA0003160078700000048
Figure GDA0003160078700000049
wherein t is2＝∑Ω[1-u(x)]m。
In order to calculate the change of the energy functional, the invention respectively calculates the change values of the regional fuzzy term and the significant perception prior term. First, the area-type blur term is rewritten into:
Figure GDA00031600787000000410
left item in the above formula
Figure GDA00031600787000000411
Can be rewritten as:
Figure GDA00031600787000000412
equation of
Figure GDA00031600787000000413
And
Figure GDA00031600787000000414
the calculation is as follows:
Figure GDA00031600787000000415
Figure GDA0003160078700000051
therefore, the temperature of the molten metal is controlled,
Figure GDA0003160078700000052
can be rewritten as
Figure GDA0003160078700000053
In the same way, calculate
Figure GDA0003160078700000054
Obtaining:
Figure GDA0003160078700000055
thus, are combined
Figure GDA0003160078700000056
And
Figure GDA0003160078700000057
calculating the area-type ambiguity term can obtain:
Figure GDA0003160078700000058
Figure GDA0003160078700000061
wherein t is1＝∑Ωu(y)mAnd t2＝∑Ω[1-u(y)]m。
The significant perceptual prior term is rewritten as:
Figure GDA0003160078700000062
using the above calculations
Figure GDA0003160078700000063
Can be obtained by
Figure GDA0003160078700000064
And
Figure GDA0003160078700000065
Figure GDA0003160078700000066
Figure GDA0003160078700000067
thus, are combined
Figure GDA0003160078700000068
And
Figure GDA0003160078700000069
calculating the significant perceptual prior term yields:
Figure GDA00031600787000000610
Figure GDA00031600787000000611
thus, by updating the constant c1、c2、s1And s2The change in energy is calculated and the membership function u (x) is updated.
Compared with the prior art, the invention has the following advantages: firstly, constructing a convex region type fuzzy item with local image information, and segmenting a gray-scale uneven image by calculating local statistical information; secondly, constructing a convexity significant perception prior fuzzy term, and improving the segmentation effect of segmenting the weak boundary image; thirdly, the change value of the convex energy functional of the segmentation model is directly calculated, and the membership function is updated according to the change value, so that the convergence rate is very high, and the segmentation result is irrelevant to the initialization condition.
Drawings
FIG. 1 is a general flow chart of a fuzzy regional active contour segmentation model based on a significant perception prior according to the present invention;
FIG. 2a illustrates various initialized contour positions according to an embodiment of the present invention;
FIG. 2b is a profile final resting position according to an embodiment of the present invention;
FIG. 3a compares the original image of the segmentation results of the present invention with other methods;
FIG. 3b is a comparison of the segmentation results of the Chan-Vese model in comparison with the segmentation results of other methods according to the present invention;
FIG. 3c compares the results of the SDREL model segmentation of the segmentation results of the present invention with that of other methods;
FIG. 3d compares the segmentation results of the present invention with segmentation results of other methods;
FIG. 3e is a group route comparing the segmentation results of the embodiment of the present invention with those of other methods.
Detailed Description
The invention relates to a fuzzy regional active contour segmentation model based on significant perception prior, which comprises the following specific implementation steps:
(1) inputting a segmentation image, setting initialization parameters: weighting coefficients lambda and alpha, maximum iteration times and an edge detection operator g;
(2) initializing level set function: target area u0(x) > 0.5, background region u0(x)＜0.5；
(3) According to the formula
Figure GDA0003160078700000071
Calculating the region saliency characteristic map and calculating the new gray value of the image
Figure GDA0003160078700000072
And
Figure GDA0003160078700000073
new pixel values of salient feature map
Figure GDA0003160078700000074
And
Figure GDA0003160078700000075
Figure GDA0003160078700000076
Figure GDA0003160078700000077
Figure GDA0003160078700000078
Figure GDA0003160078700000079
wherein t is1＝∑Ω[u(x)]mAnd t2＝∑Ω[1-u(y)]m。
(4) Calculating membership function un(x) And updating the corresponding new gray scale value
Figure GDA00031600787000000710
And
Figure GDA00031600787000000711
new pixel values of salient feature map
Figure GDA00031600787000000712
And
Figure GDA00031600787000000713
(5) calculating the energy variation difference value of a certain pixel point P according to the following formula,
Figure GDA0003160078700000081
if Δ E > 0, use unValue instead of u0Otherwise, u is maintained0The original value is unchanged;
(6) repeatedly calculating the energy change values of all pixels by using Jacobi iteration, and finishing the iteration process;
(7) and (5) repeating the steps (4) to (6) until the cycle is ended.
FIG. 2 shows the segmentation results of the natural image corresponding to the segmentation model at different initialization positions in the embodiment of the present invention as shown in FIG. 1: (a) four different initialization positions of the pseudo-zero level set function; (b) the corresponding four pseudo zero level sets final parking positions.
FIG. 3 shows a comparison of the segmentation results of the embodiment of the present invention with other methods, as shown in FIG. 3. Wherein, the graph (a) is the original images of four different flowers; the graphs (b) and (d) are respectively a Chan-Vese model, an SDREL model and the segmentation result of the embodiment of the invention; FIG. g shows a ground route.
The experimental result shows that the embodiment of the invention is irrelevant to the position of the initial contour curve, and has the characteristics of high segmentation precision, high speed and robustness to image noise.

Claims (2)

1.一种基于显著感知先验的模糊区域型活动轮廓分割模型，其能量泛函由区域型模糊项和显著感知先验项构成；区域性模糊项Eimg定义如下：1. A fuzzy regional active contour segmentation model based on salient perception prior, its energy functional is composed of regional fuzzy term and salient perception prior; the regional fuzzy term E img is defined as follows: Eimg(u)＝λ∫Ω[u(x)]m(S(x)-c1)2dx+λ∫Ω[1-u(x)]m(S(x)-c2)2dxE img (u)=λ∫ Ω [u(x)] m (S(x)-c 1 ) 2 dx+λ∫ Ω [1-u(x)] m (S(x)-c 2 ) 2 dx 其中，λ是大于0的权重系数，Ω为图像域，c1和c2是图像像素均值，u(x)是模糊成员函数，m是权重幂指数，可取值为1或2，S(x)为局部区域图；为简化计算，假设I(x)是输入图像，
Figure FDA0003184524150000011
为尺度为r1方差σ1的图像高斯卷积特征图，其中
Figure FDA0003184524150000012
为尺度为r1方差σ1的图像高斯函数；图像灰度值c1和c2定义如下：Among them, λ is the weight coefficient greater than 0, Ω is the image domain, c 1 and c 2 are the mean values of image pixels, u(x) is the fuzzy member function, m is the weight power exponent, which can be 1 or 2, S( x) is the local area map; to simplify the calculation, suppose I(x) is the input image,
Figure FDA0003184524150000011
is the image Gaussian convolution feature map with scale r 1 variance σ 1 , where
Figure FDA0003184524150000012
is an image Gaussian function with a scale of r 1 and a variance σ 1 ; the image gray values c 1 and c 2 are defined as follows:
Figure FDA0003184524150000013
Figure FDA0003184524150000013
 显著感知先验项定义如下：The salient perceptual priors are defined as follows: Esal(u)＝α∫Ωg(h(x)-s1)2·[u(x)]mdx+α∫Ωg(h(x)-s2)2·[1-u(x)]mdxE sal (u)=α∫ Ω g(h(x)-s 1 ) 2 ·[u(x)] m dx+α∫ Ω g(h(x)-s 2 ) 2 ·[1-u( x)] m dx 其中，α是大于0的权重系数，s1和s2是显著特征图像素均值，h(x)为显著性检测函数，其定义为
Figure FDA0003184524150000014
Iu是图像的区域特征均值，
Figure FDA0003184524150000015
是尺度为r2方差σ2的高斯核函数，g是边缘检测算子，其算子定义如下：where α is a weight coefficient greater than 0, s 1 and s 2 are the pixel mean values of the saliency map, and h(x) is the saliency detection function, which is defined as
Figure FDA0003184524150000014
I u is the regional feature mean of the image,
Figure FDA0003184524150000015
is a Gaussian kernel function with scale r 2 variance σ 2 , g is an edge detection operator, and its operator is defined as follows:
Figure FDA0003184524150000016
Figure FDA0003184524150000016
 其中，g是奇数行为N的方阵，▽di为当前点到邻近点的距离，K是矩阵元素个数；显著特征图像素均值s1和s2定义为：Among them, g is a square matrix with odd row N, ▽di is the distance from the current point to the adjacent point, and K is the number of matrix elements; the pixel mean values s 1 and s 2 of the salient feature map are defined as:
Figure FDA0003184524150000017
Figure FDA0003184524150000017
 因此，分割模型的能量函数表达式如下：Therefore, the energy function expression of the segmentation model is as follows:
Figure FDA0003184524150000021
Figure FDA0003184524150000021
 2.根据权利要求1所述的基于显著感知先验的模糊区域型活动轮廓分割模型，为了计算能量泛函的差值ΔE，假设P是为图像中某一像素点，对应的局部特征图灰度值为s0和隶属度为u0，对应的图像灰度值为h0；相应地，对同一固定点P的新隶属度为un，能量泛函的变化值ΔE计算公式如下：2. The fuzzy region-type active contour segmentation model based on significant perception prior according to claim 1, in order to calculate the difference ΔE of the energy functional, it is assumed that P is a certain pixel in the image, and the corresponding local feature map is gray. The degree value is s 0 and the membership degree is u 0 , and the corresponding image gray value is h 0 ; correspondingly, the new membership degree for the same fixed point P is u n , and the calculation formula of the energy functional change value ΔE is as follows:
Figure FDA0003184524150000022
Figure FDA0003184524150000022
 其中t1＝∑Ω[u(x)]m和t2＝∑Ω[1-u(x)]m，λ和α为大于0的权重系数，g是边缘检测算子，新灰度值
Figure FDA0003184524150000023
和
Figure FDA0003184524150000024
显著特征图的新像素值
Figure FDA0003184524150000025
和
Figure FDA0003184524150000026
以及un(x)定义如下：where t 1 =∑ Ω [u(x)] m and t 2 =∑ Ω [1-u(x)] m , λ and α are weight coefficients greater than 0, g is the edge detection operator, the new gray value
Figure FDA0003184524150000023
and
Figure FDA0003184524150000024
New pixel values for salient feature maps
Figure FDA0003184524150000025
and
Figure FDA0003184524150000026
and u n (x) is defined as follows:
Figure FDA0003184524150000027
Figure FDA0003184524150000027
Figure FDA0003184524150000028
Figure FDA0003184524150000028
Figure FDA0003184524150000029
Figure FDA0003184524150000029
Figure FDA00031845241500000210
Figure FDA00031845241500000210
Figure FDA00031845241500000211
Figure FDA00031845241500000211
 具体实施步骤如下：The specific implementation steps are as follows: (1)输入分割图像，设置初始化参数：权重系数λ和α，最大迭代次数，边缘检测算子g；(1) Input the segmented image and set the initialization parameters: weight coefficients λ and α, maximum number of iterations, edge detection operator g; (2)初始化水平集函数：目标区域u0(x)＞0.5，背景区域u0(x)＜0.5；(2) Initialization level set function: target area u 0 (x)>0.5, background area u 0 (x)<0.5; (3)根据公式
Figure FDA00031845241500000212
计算区域显著性特征图，并计算图像的新灰度值
Figure FDA0003184524150000031
和
Figure FDA0003184524150000032
显著特征图的新像素值
Figure FDA0003184524150000033
和
Figure FDA0003184524150000034
(3) According to the formula
Figure FDA00031845241500000212
Calculate the regional saliency feature map, and calculate the new gray value of the image
Figure FDA0003184524150000031
and
Figure FDA0003184524150000032
New pixel values for salient feature maps
Figure FDA0003184524150000033
and
Figure FDA0003184524150000034
Figure FDA0003184524150000035
Figure FDA0003184524150000035
Figure FDA0003184524150000036
Figure FDA0003184524150000036
 其中t1＝∑Ω[u(x)]m和t2＝∑Ω[1-u(x)]m；where t 1 = ∑Ω [u(x)] m and t2 = ∑Ω [1-u(x)] m ; (4)计算隶属度函数un(x)，并更新对应的新灰度值
Figure FDA0003184524150000037
和
Figure FDA0003184524150000038
显著特征图的新像素值
Figure FDA0003184524150000039
和
Figure FDA00031845241500000310
(4) Calculate the membership function u n (x), and update the corresponding new gray value
Figure FDA0003184524150000037
and
Figure FDA0003184524150000038
New pixel values for salient feature maps
Figure FDA0003184524150000039
and
Figure FDA00031845241500000310
 (5)根据如下公式计算某一像素点P能量变化差值，如果ΔE＞0，用un值代替u0，否则保持u0原始值不变：(5) Calculate the energy change difference value of a certain pixel point P according to the following formula. If ΔE>0, replace u 0 with the u n value, otherwise keep the original value of u 0 unchanged:
Figure FDA00031845241500000311
Figure FDA00031845241500000311
 其中，λ和α为大于0的权重系数，g是边缘检测算子；Among them, λ and α are weight coefficients greater than 0, and g is the edge detection operator; (6)用Jacobi迭代重复计算所有像素的能量变化值，一次迭代过程结束；(6) Repeatedly calculate the energy change value of all pixels with Jacobi iteration, and one iteration process ends; (7)重复步骤(4)-(6)直至循环结束。(7) Repeat steps (4)-(6) until the cycle ends.
CN201810630459.8A 2018-06-19 2018-06-19 An Active Contour Segmentation Model for Blurred Regions Based on Salient Perceptual Priors Active CN108898611B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810630459.8A CN108898611B (en) 2018-06-19 2018-06-19 An Active Contour Segmentation Model for Blurred Regions Based on Salient Perceptual Priors

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810630459.8A CN108898611B (en) 2018-06-19 2018-06-19 An Active Contour Segmentation Model for Blurred Regions Based on Salient Perceptual Priors

Publications (2)

Publication Number Publication Date
CN108898611A CN108898611A (en) 2018-11-27
CN108898611B true CN108898611B (en) 2021-09-24

Family

ID=64345514

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810630459.8A Active CN108898611B (en) 2018-06-19 2018-06-19 An Active Contour Segmentation Model for Blurred Regions Based on Salient Perceptual Priors

Country Status (1)

Country Link
CN (1) CN108898611B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110120057B (en) * 2019-04-16 2023-09-26 东华理工大学 Fuzzy regional active contour segmentation model based on weighted global and local fitting energy
CN110176021B (en) * 2019-05-21 2021-04-16 山东大学 Level set image segmentation method and system for saliency information combined with brightness correction
CN112102243B (en) * 2020-08-13 2023-06-09 哈尔滨工业大学（深圳） Active contour segmentation method and system combining general energy function and priori information item

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7693643B2 (en) * 2005-02-14 2010-04-06 Honeywell International Inc. Fault detection system and method for turbine engine fuel systems
US8390704B2 (en) * 2009-10-16 2013-03-05 Eastman Kodak Company Image deblurring using a spatial image prior
DE212011100130U1 (en) * 2010-08-05 2013-06-24 Philips Intellectual Property & Standards Gmbh At the same level and interactive surface mesh adaptation
CN103903251B (en) * 2012-12-30 2017-03-29 南京理工大学 Night vision image method for extracting remarkable configuration based on non-classical receptive field complex modulated
CN103366353A (en) * 2013-05-08 2013-10-23 北京大学深圳研究生院 Infrared image and visible-light image fusion method based on saliency region segmentation
CN104835168B (en) * 2015-05-15 2017-08-18 东华理工大学 Quick multiphase image dividing method based on global convex optimization Variation Model
CN106803260B (en) * 2016-12-28 2019-08-09 辽宁师范大学 Active Contour Segmentation Method of Infrared Ship Image Based on Local Entropy Convex Optimization
CN107330897B (en) * 2017-06-01 2020-09-04 福建师范大学 Image segmentation method and system thereof
CN107341800B (en) * 2017-07-10 2019-10-11 西安电子科技大学 SAR image change detection method based on superpixel saliency analysis

Also Published As

Publication number Publication date
CN108898611A (en) 2018-11-27

Similar Documents

Publication Publication Date Title
CN113052263B (en) A small sample image classification method based on manifold learning and high-order graph neural network
CN110120057B (en) Fuzzy regional active contour segmentation model based on weighted global and local fitting energy
CN109741341B (en) Image segmentation method based on super-pixel and long-and-short-term memory network
CN108776975B (en) A Visual Tracking Method Based on Joint Learning of Semi-supervised Features and Filters
Li et al. Minimization of region-scalable fitting energy for image segmentation
CN109472792B (en) An Image Segmentation Method Combining Local Energy Functional with Local Entropy and Non-convex Regular Term
CN103353987B (en) A kind of superpixel segmentation method based on fuzzy theory
CN109191477B (en) A Fuzzy Region Active Contour Segmentation Model Based on Global and Local Fitting Energy
CN108898611B (en) An Active Contour Segmentation Model for Blurred Regions Based on Salient Perceptual Priors
CN106570867A (en) ACM (Active Contour Model) image rapid segmentation method based on gray scale morphological energy method
CN105678747B (en) A kind of tooth mesh model automatic division method based on principal curvatures
CN102354396A (en) Method for segmenting image with non-uniform gray scale based on level set function
CN108876769B (en) A method for segmentation of left atrial appendage CT images
CN104616308A (en) Multiscale level set image segmenting method based on kernel fuzzy clustering
CN111047603B (en) Aerial image hybrid segmentation algorithm based on novel Markov random field and region combination
CN104835168B (en) Quick multiphase image dividing method based on global convex optimization Variation Model
Yuan et al. Hybrid method combining superpixel, random walk and active contour model for fast and accurate liver segmentation
CN110084136A (en) Context based on super-pixel CRF model optimizes indoor scene semanteme marking method
CN111145142B (en) Gray-scale uneven cyst image segmentation method based on level set algorithm
CN113838066A (en) Color image segmentation method based on improved fuzzy c-means clustering algorithm
CN111145179B (en) Gray-scale uneven image segmentation method based on level set
CN114862891B (en) Salient object detection method based on heuristic boundary optimization
CN108596926A (en) Gray threshold acquisition based on chiasma type particle cluster algorithm, method for detecting image edge
CN110517271B (en) Image level set segmentation method based on prior shape constraint
CN110738680A (en) Self-adaptive active contour segmentation model based on local energy pressure driving

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
OL01 Intention to license declared
OL01 Intention to license declared
EE01 Entry into force of recordation of patent licensing contract
EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20181127

Assignee: Anhui Yuanchen Information Technology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980005348

Denomination of invention: A Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20240507

EE01 Entry into force of recordation of patent licensing contract
EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20181127

Assignee: Anhui Tulian Information Technology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980006368

Denomination of invention: A Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20240529

Application publication date: 20181127

Assignee: Anhui Pusi Information Technology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980006339

Denomination of invention: A Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20240528

EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20181127

Assignee: JIANGXI MINGZHENG INTELLIGENT ELECTRICAL Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980017783

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241010

Application publication date: 20181127

Assignee: Jiangxi Jinyuan special equipment Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980017322

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20240930

EE01 Entry into force of recordation of patent licensing contract
EE01 Entry into force of recordation of patent licensing contract
EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20181127

Assignee: Hefei Xinguan Semiconductor Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980018198

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241011

Application publication date: 20181127

Assignee: Jiangxi Yueyin Acoustics Technology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980018134

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241011

Application publication date: 20181127

Assignee: JIANGXI TONGDA INDUSTRIAL Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980018093

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241011

Application publication date: 20181127

Assignee: JIANGXI TENGDE INDUSTRIAL Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980018090

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241011

Application publication date: 20181127

Assignee: FUZHOU ZHONGYUAN INDUSTRIAL Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980018085

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241011

EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20181127

Assignee: JIANGXI TEXIN INDUSTRIAL Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980036102

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241211

Application publication date: 20181127

Assignee: Shangde United Electric Group Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980036099

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241211

Application publication date: 20181127

Assignee: JIANGXI QILI INDUSTRIAL DEVELOPMENT Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980036093

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241211

Application publication date: 20181127

Assignee: Pengsheng Construction Group Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980036089

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241211

Application publication date: 20181127

Assignee: Jiangxi laibo'er laboratory equipment manufacturing Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980036083

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241211

Application publication date: 20181127

Assignee: JIANGXI JINRUI ENVIRONMENTAL PROTECTION TECHNOLOGY Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980036076

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241211

Application publication date: 20181127

Assignee: JIANGXI DONGYAN PHARMACEUTICAL Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980036057

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241211

Application publication date: 20181127

Assignee: Jiangxi Bond Technology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980036049

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241211

EE01 Entry into force of recordation of patent licensing contract
EE01 Entry into force of recordation of patent licensing contract
EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20181127

Assignee: JIANGXI XINDING TECHNOLOGY Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980038099

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241218

Application publication date: 20181127

Assignee: Jiangxi Saili Medical Equipment Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980038093

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241218

Application publication date: 20181127

Assignee: Jiangxi Aimei Technology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980038077

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241217

Application publication date: 20181127

Assignee: Jiangxi Guowei photoelectric display Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980038067

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241217

Application publication date: 20181127

Assignee: Jiangxi Qiannong Biotechnology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980038026

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241217

Application publication date: 20181127

Assignee: JIANGXI RUIYI YUNCHENG TECHNOLOGY CO.,LTD.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980037922

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241217

Application publication date: 20181127

Assignee: JIANGXI RUIBOTE BIOTECHNOLOGY CO.,LTD.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980037883

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241217

Application publication date: 20181127

Assignee: Guangchang Minnong Technology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980037870

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241217

Application publication date: 20181127

Assignee: Shangrao Aichi Biotechnology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980037837

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241217

EE01 Entry into force of recordation of patent licensing contract
EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20181127

Assignee: Jiangxi Ruoshan Technology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980040209

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241219

Application publication date: 20181127

Assignee: JIANGXI ZHANBANG TECHNOLOGY Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980040077

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241219

Application publication date: 20181127

Assignee: Shangrao Chuchu Electric Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980039831

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241219

EE01 Entry into force of recordation of patent licensing contract
EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20181127

Assignee: Jiangxi Tongtong Photoelectric Technology Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2024980041892

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20241224

EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20181127

Assignee: Jiangxi Yulin Industrial Co.,Ltd.

Assignor: EAST CHINA INSTITUTE OF TECHNOLOGY

Contract record no.: X2025980000665

Denomination of invention: Fuzzy Region Active Contour Segmentation Model Based on Significant Perception Prior

Granted publication date: 20210924

License type: Open License

Record date: 20250107

EE01 Entry into force of recordation of patent licensing contract
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