CN112102243B - Active contour segmentation method and system combining general energy function and priori information item - Google Patents
Active contour segmentation method and system combining general energy function and priori information item Download PDFInfo
- Publication number
- CN112102243B CN112102243B CN202010811710.8A CN202010811710A CN112102243B CN 112102243 B CN112102243 B CN 112102243B CN 202010811710 A CN202010811710 A CN 202010811710A CN 112102243 B CN112102243 B CN 112102243B
- Authority
- CN
- China
- Prior art keywords
- energy function
- priori information
- image
- function
- segmentation
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a method and a system for segmenting an active contour by combining a general energy function and a priori information item, wherein the method for segmenting the active contour comprises the following steps: step 1: selecting an acquisition mode of priori information to obtain a segmentation result of fuzzy clustering; step 2: inputting an image to be segmented, an initial level set function, a stopping condition, an initial Bregman variable and an initial auxiliary variable; step 3: calculating the approximate intensities of the images inside and outside the initial contour line; step 4: calculating an image data fitting item; step 5: calculating a level set function; step 6: calculating an auxiliary vector and Bregman variables; step 7: and (5) iterating. The beneficial effects of the invention are as follows: 1. the invention uses the priori information, so that the accuracy of the result is greatly improved, and the automatically acquired priori information can effectively reduce manual operation and improve the working efficiency.
Description
Technical Field
The invention relates to the field of image segmentation and processing, in particular to a method and a system for segmenting an active contour by combining a general energy function and a priori information item.
Background
The image segmentation has a wide application foundation in the fields of industrial production and medical folk life, can effectively improve the level of industrial automation, and realizes the positioning and navigation of targets in the traffic field. In the medical imaging field, doctors are assisted in performing surgery and locating a lesion area.
In the field of medical image processing, an image segmentation algorithm effectively segments an ultrasonic image and a nuclear magnetic resonance image, can distinguish different organs and tissues, marks names, assists doctors in diagnosing related diseases, improves the diagnosis efficiency of medical departments, and popularizes experience and practice to a wider basic medical system. In the traffic field, the image segmentation technology is applied to the automatic driving field, road obstacles are identified, and driving safety and efficiency are improved.
Currently, existing image segmentation methods can be divided into two types, namely a traditional method based on a boundary detection method, a method based on region growth, an active contour model and the like, and a latest machine learning method.
The boundary detection method searches for a position of discontinuous pixel values at the edge through a boundary detection operator, and is called as a 'sudden change'. Such discontinuities may be detected by deriving derivatives, the positions of which correspond to the extreme points of the first derivative and to the zero points of the second derivative for step-like edges. Representative operators are a LoG operator, a Canny operator, and the like.
The region growing method combines adjacent pixels or regions with similar growth point properties with the growth point according to the pixel value to form a new growth point, and repeats the process until growth cannot be performed.
The RSF model commonly used in active contour models is also a region segmentation based model. The method sets omega as the image definition field,
is a gray image, f 1 ,f 2 Is a function of two representing the image intensities inside and outside the segmentation boundary. The task of this model is to minimize the energy function to find an approximation image u and its boundary C, thus dividing the image into two parts.
The latest deep learning method has been widely used in the field of image segmentation since the advent of the deep learning method. Different networks may handle different segmentation problems. Earlier image segmentation methods have more involved the discovery of edges (lines and curves) or gradient elements. The concept of semantic segmentation will now allow a computer to provide pixel-level image understanding in a purely human-perceived manner. Semantic segmentation aims at solving the problem by gathering image parts belonging to the same object together, thereby expanding the application field of the semantic segmentation.
R-CNN is the first algorithm to apply deep neural networks to target detection, but it is inefficient, which the latter has made many improvements. For example, fully considering detail information, use median rebalancing weights for class prediction, and so on.
FCN is an upsampling-based method, in which the picture is upsampled (pixel-enlarged) and then convolved in an upsampling structure. This reduces the loss of image detail information in the convolution. U-NET is a network structure designed for medical image segmentation. Aiming at the specificity of medical images, the method proposed by the U-NET effectively improves the training and detecting effect by using a small amount of data sets, and simultaneously provides an effective method for processing large-size images, thereby being a network structure which is widely used at present.
The mathematical model of the boundary detection algorithm is simple and popular and easy to understand, but the segmentation result is poor, only partial contour boundaries are found, the boundary detection algorithm is easy to be interfered by other areas in the image, and the segmentation result in the background exists outside a target area of a plurality of objects. The limitation of the area growth method is the selection of the pixels of the initial growth points and the formulation of the growth criteria. These two points will have a great impact on the segmentation results.
The RSF model is more sensitive to initialization, and in particular, the initial contour and parameters in the model have a greater impact on the results. Therefore, a good result is obtained by utilizing the RSF model, an initial contour is required to be selected according to different image characteristics, and the parameters are adjusted through multiple tests, so that the method is labor-consuming and labor-consuming, and has a large lifting space.
As an image segmentation algorithm which is widely used at present, the deep learning method still has a plurality of defects. Firstly, the training model needs massive data, the medical image is different from the natural image, the medical image is extremely difficult to obtain, and the label corresponding to the image, namely the lesion position, needs to be marked by a professional doctor, so that the scarcity of medical image resources is reflected. Second, the deep learning algorithm is highly dependent on the computer hardware facilities, which will bring about a huge cost investment. Furthermore, the specificity of medical images places higher demands on the configuration of the computer. If image quality is reduced, the calculation efficiency is pursued, which does not play a role in accurately diagnosing diseases.
Disclosure of Invention
The invention provides an active contour segmentation method combining a general energy function and a priori information item, which comprises the following steps:
step 1: selecting the acquisition mode of priori information to obtain the segmentation result of fuzzy clustering, and taking the segmentation result as an input ψ spi The method comprises the steps of carrying out a first treatment on the surface of the Selecting an energy function as H gen (Ψ)；
Step 2: inputting an image I to be segmented, and an initial level set function ψ 0 Stop condition γ, initial Bregman variable b 0 =0 and initial auxiliary variable h 0 ＝0；
Step 3: calculating approximate intensities of images inside and outside an initial contour line
Step 4: calculating an image data fitting term S k ；
Step 5: calculating a level set function ψ k+1 ；
Step 6: calculate the auxiliary vector h k+1 And Bregman variable b k+1 ；
Step 7: iteration
Step 8: judging |ψ k+1 -Ψ k The I is less than or equal to gamma; if ||ψ k+1 -Ψ k Step 9 is executed if the level is less than or equal to gamma, otherwise, step 4 is returned, and the image data fitting item S is continuously calculated k ；
Step 9: outputting the final result, namely the segmentation contour line Z k+1 。
As a further improvement of the present invention, in the step 1, the following steps are further performed:
step S1: establishing a general image segmentation energy function based on priori information;
step S2, defining prior information items and selecting energy function items;
and S3, selecting specific priori information, and minimizing the energy functional.
As a further improvement of the present invention, in the step S1, specifically, it includes:
consider an image
Is a level set function over definition domain D, we define an energy function H (ψ):
wherein H is gen (ψ) refers to a general energy function, we choose a boundary-based energy function, a region-based energy function or other type of energy function, H, based on the characteristics and needs of the different images spi (ψ) refers to a general a priori information item.
As a further improvement of the present invention, in the step S2, specifically, it includes:
defining a priori information item and selecting an energy function item, the priori information item H spi (ψ) we define as:
H spi (Ψ)＝∫ D |Ψ(x)-Ψ spi (x)| 2 dx, (2)
wherein ψ is spi For the energy function term, we choose to take a region-based energy function as an example, and the energy function is expressed as an inner product:
s is the image data fitting term and ω is a boundary detection operator, defined as:
in the formula, x and y represent the positions of pixel points in the image, lambda 1 ,λ 2 Beta is a parameter greater than zero, G σ Is a Gaussian kernel function, σ is the standard deviation, v 1 ,v 2 Respectively are movableApproximate image intensities inside and outside the contour line.
As a further improvement of the present invention, in the step S3, specifically, it includes:
according to formulas (2) and (3), the minimization problem of the model is:
solving this minimization problem, first, using the introduced Bregman variable b= (b) x ,b y ) And an auxiliary variable h= (h x ,h y ) So that
The minimization problem equation (5) is solved by the following equation:
wherein, psi is k+1 Is a level set function, h k+1 Is an auxiliary variable, b k+1 And b k Is a Bregman variable;
finally, we denote the evolving active contour by z, defined as:
z k+1 ＝{x:Ψ k+1 (x)＝0}. (7)
the invention also discloses a movable contour segmentation system combining the general energy function and the prior information item, which comprises:
a selection unit: obtaining a segmentation result of fuzzy clustering by selecting an acquisition mode of priori information, wherein the segmentation result is used as an input ψ spi The method comprises the steps of carrying out a first treatment on the surface of the Selecting an energy function as H gen (Ψ)；
An input unit: for inputting the image I to be segmented, an initial level set function ψ 0 Stop condition γ, initial Bregman variable b 0 =0 and initial auxiliary variable h 0 ＝0；
A first calculation unit: for calculating approximate intensities of images inside and outside an initial contour line
/>
A second calculation unit: for calculating image data fitting terms S k ；
A third calculation unit: for calculating level set function ψ k+1 ；
A fourth calculation unit: for calculating auxiliary vector h k+1 And Bregman variable b k+1 ；
Iteration unit: for iteration
A judging unit: for judging ψ k+1 -Ψ k The I is less than or equal to gamma; if ||ψ k+1 -Ψ k Step 9 is executed if the level is less than or equal to gamma, otherwise, step 4 is returned, and the image data fitting item S is continuously calculated k ；
An output unit: for outputting the final result segmentation contour z k+1 。
As a further improvement of the present invention, in the selecting unit, further comprising:
and (3) a building module: for creating a generic a priori information based image segmentation energy function;
the definition module: for defining a priori information items and selecting energy function items;
and (3) a minimization module: for selecting specific prior information, minimizing the energy functional.
As a further improvement of the present invention, in the building module, specifically, it includes:
consider an image
Is a level set function over definition domain D, we define an energy function H (ψ):
wherein H is gen (ψ) refers to a general energy function, we choose a boundary-based energy function, a region-based energy function or other type of energy function, H, based on the characteristics and needs of the different images spi (ψ) refers to a general a priori information item.
As a further improvement of the present invention, in the definition module, specifically includes:
defining a priori information item and selecting an energy function item, the priori information item H spi (ψ) we define as:
H spi (Ψ)＝∫ D |Ψ(x)-Ψ spi (x)| 2 dx, (2)
wherein ψ is spi For the energy function term, we choose to take a region-based energy function as an example, and the energy function is expressed as an inner product:
s is the image data fitting term and ω is a boundary detection operator, defined as:
in the formula, x and y represent the positions of pixel points in the image, lambda 1 ,λ 2 Beta is a parameter greater than zero, G σ Is a Gaussian kernel function, σ is the standard deviation, v 1 ,v 2 Respectively the approximate image intensities inside and outside the active contour line.
As a further improvement of the present invention, in the minimization module, specifically, it includes:
according to formulas (2) and (3), the minimization problem of the model is:
solving this minimization problem, first, using the introduced Bregman variable b= (b) x ,b y ) And an auxiliary variable h= (h x ,h y ) So that
The minimization problem equation (5) is solved by the following equation:
wherein, psi is k+1 Is a level set function, h k+1 Is an auxiliary variable, b k+1 And b k Is a Bregman variable;
finally, we denote the evolving active contour by z, defined as:
z k+1 ＝{x:Ψ k+1 (x)＝0}. (7)
the beneficial effects of the invention are as follows: 1. the movable contour segmentation method uses the priori information, so that the accuracy of a result is greatly improved, and in addition, the priori information automatically acquired by the movable contour segmentation method can effectively reduce manual operation and improve the working efficiency; 2. as the general energy function is obtained from various existing mature algorithms, the active contour segmentation method not only increases the applicability to different types of images, but also increases the output of segmentation results; 3. the method for segmenting the movable contour solves the minimized energy functional by using a split Bregman method, and compared with the existing gradient descent method, the method is robust to initialization; 4. the energy function used by the active contour segmentation method is a convex function, so that the local optimal solution is not involved in the process of minimizing the energy function, the segmentation result is accurate, and meanwhile, the speed of minimizing the energy functional by using the split Bregman method is faster than that of the traditional gradient descent method under the condition of limited iterations, so that the efficiency of the active contour segmentation method is greatly improved.
Drawings
FIG. 1 is a flow chart of an active contour segmentation method of the present invention;
fig. 2 is a flow chart of the selection of a priori information of the present invention.
Detailed Description
As shown in fig. 1, the invention discloses an active contour segmentation method combining a general energy function and a priori information items, which comprises the following steps:
step 1: selecting an acquisition mode of priori information, taking fuzzy clustering as the priori information as an example, and obtaining a segmentation result of the fuzzy clustering as an input ψ spi The method comprises the steps of carrying out a first treatment on the surface of the Selecting an energy function as H gen (Ψ)；
Step 2: inputting an image I to be segmented, and an initial level set function ψ 0 Stop condition γ, initial Bregman variable b 0 =0 and initial auxiliary variable h 0 ＝0；
Step 3: calculating approximate intensities of images inside and outside an initial contour line
Step 4: calculating an image data fitting term S k ；
Step 5: calculating a level set function ψ k+1 ；
Step 6: calculate the auxiliary vector h k+1 And Bregman variable b k+1 ；
Step 7: iteration
Step 8: judging |ψ k+1 -Ψ k The I is less than or equal to gamma; if ||ψ k+1 -Ψ k Step 9 is executed if the level is less than or equal to gamma, otherwise, step 4 is returned, and the image data fitting item S is continuously calculated k Until the judgment formula of the step 8 is established;
step 9: outputting the final result segmentation contour line Z k+1 。
The invention discloses a movable contour segmentation method, which is a quick algorithm for providing a series of flexible medical image disease diagnosis and treatment and natural image target recognition segmentation, and aims to solve various image segmentation problems and effectively improve segmentation accuracy, and as shown in fig. 2, in the step 1, the method further comprises the following steps:
step S1: establishing a general image segmentation energy function based on priori information;
in the step S1, specifically, the method includes:
consider an image
Is a level set function over definition domain D, we define an energy function H (ψ):
wherein H is gen (ψ) refers to a general energy function, which we can choose boundary-based energy functions, region-based energy functions or other types of energy functions, H, according to the characteristics and needs of the different images spi (ψ) refers to a general item of a priori information, which is related to both the morphology of the object itself and the subjective experience of the person, and thus can be obtained automatically or manually. Step S2, defining prior information items and selecting energy function items, taking an energy function based on a region as an example;
in the step S2, specifically, the method includes:
defining a priori information item and selecting an energy function item, taking a region-based energy function as an example, the a priori information item H spi (ψ) we define as:
H spi (Ψ)＝∫ D |Ψ(x)-Ψ spi (x)| 2 dx, (2)
wherein ψ is spi Level set function representing the result of a priori information segmentation, ψ represents the evolving level set function, and for the energy function term we choose to use a region-based energy functionThe number is for example, and the energy function is expressed in terms of an inner product as:
s is the image data fitting term and ω is a boundary detection operator, defined as:
in the formula, x and y represent the positions of pixel points in the image, lambda 1 ,λ 2 Beta is a parameter greater than zero, G σ Is a Gaussian kernel function, σ is the standard deviation, v 1 ,v 2 Respectively the approximate image intensities inside and outside the active contour line. And S3, selecting specific priori information, taking fuzzy clustering as an example of the priori information, and minimizing the energy functional.
In the step S3, the method specifically includes:
according to formulas (2) and (3), the minimization problem of the model is:
solving this minimization problem, first, using the introduced Bregman variable b= (b) x ,b y ) And an auxiliary variable h= (h x ,h y ) So that
The minimization problem equation (5) can be solved by the following equation: />
Wherein, psi is k+1 Is a level set function, h k+1 Is an auxiliary variable, b k+1 And b k Is a Bregman variable;
finally, we denote the evolving active contour by z, defined as:
z k+1 ＝{x:Ψ k+1 (x)＝0}. (7)
the segmentation result and model evaluation, namely the active contour segmentation method combining the general energy function and the prior information item can segment partial medical images, including melanoma images, heart images, human brain images and the like, and can be used for auxiliary diagnosis of related diseases. The active contour segmentation method is mainly used for comparing and analyzing segmentation results of the traditional segmentation model and the deep learning model. On the index of quantitative evaluation, we select the F measure as the index for measuring the quality of the segmentation result. Taking melanoma image datasets as an example, we have improved F values by 81.20%,18.93%,8.16% and 5.02% compared to RSF, CV, FCM and U-Net methods, respectively. In time, the active contour segmentation method is stable in time consumption and free from parameter disturbance.
The invention also discloses a movable contour segmentation system combining the general energy function and the prior information item, which comprises:
a selection unit: obtaining a segmentation result of fuzzy clustering by selecting an acquisition mode of priori information, wherein the segmentation result is used as an input ψ spi The method comprises the steps of carrying out a first treatment on the surface of the Selecting an energy function as H gen (Ψ)；
An input unit: for inputting the image I to be segmented, an initial level set function ψ 0 Stop condition γ, initial Bregman variable b 0 =0 and initial auxiliary variable h 0 ＝0；
A first calculation unit: for calculating approximate intensities of images inside and outside an initial contour line
A second calculation unit: for calculating image data fitting terms S k ；
A third calculation unit: for calculating level set function ψ k+1 ；
A fourth calculation unit: for calculating auxiliary vectorsh k+1 And Bregman variable b k+1 ；
Iteration unit: for iteration
A judging unit: for judging ψ k+1 -Ψ k The I is less than or equal to gamma; if ||ψ k+1 -Ψ k Step 9 is executed if the level is less than or equal to gamma, otherwise, step 4 is returned, and the image data fitting item S is continuously calculated k ；
An output unit: for outputting the final result segmentation contour Z k+1 z k+1 。
In the selecting unit, further comprising:
and (3) a building module: for creating a generic a priori information based image segmentation energy function;
the definition module: for defining a priori information items and selecting energy function items;
and (3) a minimization module: for selecting specific prior information, minimizing the energy functional.
The establishing module specifically comprises the following steps:
consider an image
Is a level set function over definition domain D, we define an energy function H (ψ):
wherein H is gen (ψ) refers to a general energy function, which we can choose boundary-based energy functions, region-based energy functions or other types of energy functions, H, according to the characteristics and needs of the different images spi (ψ) refers to a general item of a priori information, which is related to both the morphology of the object itself and the subjective experience of the person, and thus can be obtained automatically or manually.
The definition module specifically comprises:
defining a priori information item and selecting an energy function item, taking a region-based energy function as an example, the a priori information item H spi (ψ) we define as:
H spi (Ψ)＝∫ D |Ψ(x)-Ψ spi (x)| 2 dx, (2)
wherein ψ is spi For the energy function term, we choose to take a region-based energy function as an example, and the energy function is expressed as an inner product:
s is the image data fitting term and ω is a boundary detection operator, defined as:
in the formula, x and y represent the positions of pixel points in the image, lambda 1 ,λ 2 Beta is a parameter greater than zero, G σ Is a Gaussian kernel function, σ is the standard deviation, v 1 ,v 2 Respectively the approximate image intensities inside and outside the active contour line.
The minimizing module specifically comprises:
according to formulas (2) and (3), the minimization problem of the model is:
solving this minimization problem, first, using the introduced Bregman variable b= (b) x ,b y ) And an auxiliary variable h= (h x ,h y ) So that
The minimization problem equation (5) can be solved by the following equation:
wherein, psi is k+1 Is a level set function, h k+1 Is an auxiliary variable, b k+1 And b k Is a Bregman variable;
finally, we denote the evolving active contour by z, defined as:
z k+1 ＝{x:Ψ k+1 (x)＝0}. (7)
the beneficial effects of the invention are as follows: 1. the movable contour segmentation method uses the priori information, so that the accuracy of a result is greatly improved, and in addition, the priori information automatically acquired by the movable contour segmentation method can effectively reduce manual operation and improve the working efficiency; 2. as the general energy function is obtained from various existing mature algorithms, the active contour segmentation method not only increases the applicability to different types of images, but also increases the output of segmentation results; 3. the active contour segmentation method of the invention solves the minimized energy functional by using the split Bregman method later, and compared with the existing gradient descent method, the algorithm is robust to initialization. 4. The energy function used by the active contour segmentation method is a convex function, so that the local optimal solution is not involved in the process of minimizing the energy function, the segmentation result is accurate, and meanwhile, the speed of minimizing the energy functional by using the split Bregman method is faster than that of the traditional gradient descent method under the condition of limited iterations, so that the efficiency of the active contour segmentation method is greatly improved.
The foregoing is a further detailed description of the invention in connection with the preferred embodiments, and it is not intended that the invention be limited to the specific embodiments described. It will be apparent to those skilled in the art that several simple deductions or substitutions may be made without departing from the spirit of the invention, and these should be considered to be within the scope of the invention.
Claims (6)
1. A method of active contour segmentation combining a generic energy function with a priori information items, comprising the steps of:
step 1: selecting the acquisition mode of priori information to obtain the segmentation result of fuzzy clustering, and taking the segmentation result as an input ψ spi The method comprises the steps of carrying out a first treatment on the surface of the Selecting an energy function as H gen (Ψ)；
Step 2: inputting an image I to be segmented, and an initial level set function ψ 0 Stop condition γ, initial Bregman variable b 0 =0 and initial auxiliary variable h 0 ＝0；
Step 3: calculating approximate intensities of images inside and outside an initial contour line
Step 4: calculating an image data fitting term S k ；
Step 5: calculating a level set function ψ k+1 ；
Step 6: calculate the auxiliary vector h k+1 And Bregman variable b k+1 ；
Step 7: iteration
Step 8: judging |ψ k+1 -Ψ k The I is less than or equal to gamma; if ||ψ k+1 -Ψ k Step 9 is executed if the level is less than or equal to gamma, otherwise, step 4 is returned, and the image data fitting item S is continuously calculated k ；
Step 9: outputting the final result, namely the segmentation contour line Z k+1 ；
In the step 1, the method further comprises the following steps:
step S1: establishing a general image segmentation energy function based on priori information;
step S2, defining prior information items and selecting energy function items;
s3, selecting specific priori information, and minimizing an energy functional;
in the step S3, the method specifically includes:
according to formulas (2) and (3), the minimization problem of the model is:
solving this minimization problem, first, using the introduced Bregman variable b= (b) x ,b y ) And an auxiliary variable h= (h x ,h y ) So that
The minimization problem equation (5) is solved by the following equation:
wherein, psi is k+1 Is a level set function, h k+1 Is an auxiliary variable, b k+1 And b k Is a Bregman variable;
finally, we denote the evolving active contour by z, defined as:
z k+1 ＝{x:Ψ k+1 (x)＝0}. (7)。
2. the active contour segmentation method according to claim 1, characterized in that in said step S1, specifically comprising:
consider an image I:
Ψ:/>
is a level set function over definition domain D, we define an energy function H (ψ):
wherein H is gen (ψ) refers to a general energy function, we choose a boundary-based energy function, a region-based energy function or other type of energy function, H, based on the characteristics and needs of the different images spi (ψ) refers to a general a priori information item.
3. The active contour segmentation method according to claim 1, characterized in that in said step S2, specifically comprising:
defining a priori information item and selecting an energy function item, the priori information item H spi (ψ) we define as:
H spi (Ψ)＝∫ D |Ψ(x)-Ψ spi (x)| 2 dx, (2)
wherein ψ is spi For the energy function term, we choose to take a region-based energy function as an example, and the energy function is expressed as an inner product:
s is the image data fitting term and ω is a boundary detection operator, defined as:
in the formula, x and y represent the positions of pixel points in the image, lambda 1 ,λ 2 Beta is a parameter greater than zero, G σ Is a Gaussian kernel function, σ is the standard deviation, v 1 ,v 2 Respectively the approximate image intensities inside and outside the active contour line.
4. An active contour segmentation system that combines a generic energy function with a prior information item, comprising:
a selection unit: obtaining a segmentation result of fuzzy clustering by selecting an acquisition mode of priori information, wherein the segmentation result is used as an input ψ spi The method comprises the steps of carrying out a first treatment on the surface of the Selecting an energy function as H gen (Ψ)；
An input unit: for inputting the image I to be segmented, an initial level set function ψ 0 Stop condition γ, initial Bregman variable b 0 =0 and initial auxiliary variable h 0 ＝0；
A first calculation unit: for calculating approximate intensities of images inside and outside an initial contour line
A second calculation unit: for calculating image data fitting terms S k ；
A third calculation unit: for calculating level set function ψ k+1 ；
A fourth calculation unit: for calculating auxiliary vector h k+1 And Bregman variable b k+1 ；
Iteration unit: for iteration
A judging unit: for judging ψ k+1 -Ψ k The I is less than or equal to gamma; if ||ψ k+1 -Ψ k Executing the output unit if the I is less than or equal to gamma, otherwise, returning to the second calculation unit, and continuing to calculate the image data fitting item S k ；
An output unit: for outputting final result segmentation contour Z k+1 ；
In the selecting unit, further comprising:
and (3) a building module: for creating a generic a priori information based image segmentation energy function;
the definition module: for defining a priori information items and selecting energy function items;
and (3) a minimization module: the method is used for selecting specific priori information and minimizing energy functional;
the minimizing module specifically comprises:
according to formulas (2) and (3), the minimization problem of the model is:
solving this minimization problem, first, using the introduced Bregman variable b= (b) x ,b y ) And an auxiliary variable h= (h x ,h y ) So that
The minimization problem equation (5) is solved by the following equation: />
Wherein, psi is k+1 Is a level set function, h k+1 Is an auxiliary variable, b k+1 And b k Is a Bregman variable;
finally, we denote the evolving active contour by z, defined as:
z k+1 ＝{x:Ψ k+1 (x)＝0}. (7)。
5. the active contour segmentation system as defined in claim 4, wherein the building module comprises:
consider an image I:
Ψ:/>
is a level set function over definition domain D, we define an energy functionH(ψ)：
Wherein H is gen (ψ) refers to a general energy function, we choose a boundary-based energy function, a region-based energy function or other type of energy function, H, based on the characteristics and needs of the different images spi (ψ) refers to a general a priori information item.
6. The active contour segmentation system as defined in claim 4, further comprising, in the definition module:
defining a priori information item and selecting an energy function item, the priori information item H spi (ψ) we define as:
H spi (Ψ)＝∫ D |Ψ(x)-Ψ spi (x)| 2 dx, (2)
wherein ψ is spi For the energy function term, we choose to take a region-based energy function as an example, and the energy function is expressed as an inner product:
s is the image data fitting term and ω is a boundary detection operator, defined as:
in the formula, x and y represent the positions of pixel points in the image, lambda 1 ,λ 2 Beta is a parameter greater than zero, G σ Is a Gaussian kernel function, σ is the standard deviation, v 1 ,v 2 Respectively the approximate image intensities inside and outside the active contour line.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010811710.8A CN112102243B (en) | 2020-08-13 | 2020-08-13 | Active contour segmentation method and system combining general energy function and priori information item |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010811710.8A CN112102243B (en) | 2020-08-13 | 2020-08-13 | Active contour segmentation method and system combining general energy function and priori information item |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112102243A CN112102243A (en) | 2020-12-18 |
| CN112102243B true CN112102243B (en) | 2023-06-09 |
Family
ID=73753625
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010811710.8A Active CN112102243B (en) | 2020-08-13 | 2020-08-13 | Active contour segmentation method and system combining general energy function and priori information item |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112102243B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106447688A (en) * | 2016-03-31 | 2017-02-22 | 大连海事大学 | Method for effectively segmenting hyperspectral oil-spill image |
| CN108898611A (en) * | 2018-06-19 | 2018-11-27 | 东华理工大学 | Fuzzy region active contour parted pattern based on significant perception priori |
| CN109191477A (en) * | 2018-09-21 | 2019-01-11 | 东华理工大学 | Fuzzy region type active contour parted pattern based on overall situation and partial situation's fitting energy |
| CN110120057A (en) * | 2019-04-16 | 2019-08-13 | 东华理工大学 | Fuzzy region sexuality contours segmentation model based on the global and local fitting energy of weight |
| CN110969635A (en) * | 2019-12-12 | 2020-04-07 | 哈尔滨工业大学（深圳） | Robust rapid image segmentation method based on prior constraint level set framework |
| CN111462145A (en) * | 2020-04-01 | 2020-07-28 | 重庆大学 | Active contour image segmentation method based on double-weight symbol pressure function |
-
2020
- 2020-08-13 CN CN202010811710.8A patent/CN112102243B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106447688A (en) * | 2016-03-31 | 2017-02-22 | 大连海事大学 | Method for effectively segmenting hyperspectral oil-spill image |
| CN108898611A (en) * | 2018-06-19 | 2018-11-27 | 东华理工大学 | Fuzzy region active contour parted pattern based on significant perception priori |
| CN109191477A (en) * | 2018-09-21 | 2019-01-11 | 东华理工大学 | Fuzzy region type active contour parted pattern based on overall situation and partial situation's fitting energy |
| CN110120057A (en) * | 2019-04-16 | 2019-08-13 | 东华理工大学 | Fuzzy region sexuality contours segmentation model based on the global and local fitting energy of weight |
| CN110969635A (en) * | 2019-12-12 | 2020-04-07 | 哈尔滨工业大学（深圳） | Robust rapid image segmentation method based on prior constraint level set framework |
| CN111462145A (en) * | 2020-04-01 | 2020-07-28 | 重庆大学 | Active contour image segmentation method based on double-weight symbol pressure function |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112102243A (en) | 2020-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Marin et al. | Obtaining optic disc center and pixel region by automatic thresholding methods on morphologically processed fundus images | |
| CN104732213B (en) | A kind of area of computer aided Mass detection method based on mammary gland magnetic resonance image | |
| Panda et al. | New binary Hausdorff symmetry measure based seeded region growing for retinal vessel segmentation | |
| Roy et al. | Blood vessel segmentation of retinal image using Clifford matched filter and Clifford convolution | |
| CN105912874A (en) | Liver three-dimensional database system constructed on the basis of DICOM (Digital Imaging and Communications in Medicine) medical image | |
| Pathan et al. | Automated detection of optic disc contours in fundus images using decision tree classifier | |
| Díaz-Pernil et al. | Fully automatized parallel segmentation of the optic disc in retinal fundus images | |
| CN104732520A (en) | Cardio-thoracic ratio measuring algorithm and system for chest digital image | |
| CN112288794B (en) | Method and device for measuring blood vessel diameter of fundus image | |
| CN109087310A (en) | Dividing method, system, storage medium and the intelligent terminal of Meibomian gland texture region | |
| US7747051B2 (en) | Distance transform based vessel detection for nodule segmentation and analysis | |
| CN113034507A (en) | CCTA image-based coronary artery three-dimensional segmentation method | |
| CN113160120A (en) | Liver blood vessel segmentation method and system based on multi-mode fusion and deep learning | |
| Jian et al. | Diagnosis of left ventricular hypertrophy using convolutional neural network | |
| Shabani et al. | An active contour model using matched filter and Hessian matrix for retinalvessels segmentation | |
| Dikkala et al. | A comprehensive analysis of morphological process dependent retinal blood vessel segmentation | |
| WO2015150320A1 (en) | Segmentation of tubular organ structures | |
| KR101251822B1 (en) | System and method for analysising perfusion in dynamic contrast-enhanced lung computed tomography images | |
| Padmasini et al. | State-of-the-art of level-set methods in segmentation and registration of spectral domain optical coherence tomographic retinal images | |
| Schenk et al. | Automatic high-speed video glottis segmentation using salient regions and 3D geodesic active contours | |
| Kanca et al. | Learning hand-crafted features for k-NN based skin disease classification | |
| CN112102243B (en) | Active contour segmentation method and system combining general energy function and priori information item | |
| CN110969635B (en) | Robust rapid image segmentation method based on priori constraint level set frame | |
| CN109816665B (en) | Rapid segmentation method and device for optical coherence tomography image | |
| CN108416792B (en) | Medical computed tomography image segmentation method based on active contour model |
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 |