WO2020233084A1 - 一种图像分割方法、装置、存储介质及终端设备 - Google Patents
一种图像分割方法、装置、存储介质及终端设备 Download PDFInfo
- Publication number
- WO2020233084A1 WO2020233084A1 PCT/CN2019/122572 CN2019122572W WO2020233084A1 WO 2020233084 A1 WO2020233084 A1 WO 2020233084A1 CN 2019122572 W CN2019122572 W CN 2019122572W WO 2020233084 A1 WO2020233084 A1 WO 2020233084A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- frog
- membership
- preset
- degree
- population
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
Definitions
- This application relates to the field of image processing technology, and in particular to an image segmentation method, device, computer-readable storage medium, and terminal equipment.
- Image segmentation refers to the technology and process of dividing an image into a number of specific regions with unique properties and proposing objects of interest.
- Fuzzy C-means clustering algorithm (FCM) is a common method of image segmentation, which is based on fuzzy mathematics.
- FCM Fuzzy C-means clustering algorithm
- K-means algorithm realizes clustering by optimizing a fuzzy objective function. It does not think that each point can only belong to a certain class like K-means clustering, but gives each point a corresponding degree of membership. The degree of membership can better describe the characteristics of edge pixels and is suitable for dealing with the inherent uncertainty of things.
- FCM clustering for image segmentation can reduce human intervention to achieve simple image segmentation, but FCM clustering greatly depends on the initial clustering center. Once the initial clustering center is determined unreasonably, it will seriously affect the image segmentation effect .
- the embodiments of the application provide an image segmentation method, device, computer readable storage medium, and terminal equipment, which can reasonably determine the initial cluster center of FCM clustering, improve image segmentation efficiency and segmentation effect, and can effectively reduce the image segmentation process Computational complexity in.
- the first aspect of the embodiments of the present application provides an image segmentation method, including:
- the preset leaping algorithm is initialized according to the number of clusters and the pixel points, and the pixel points are optimized through the preset leaping algorithm after initialization to obtain the local optimum of the original image pixel;
- an image segmentation device including:
- the original image acquisition module is used to acquire the original image to be segmented and determine the number of clusters and pixels of the original image
- the optimization processing module is configured to initialize a preset frog leaping algorithm according to the number of clusters and the pixel points, and perform optimization processing on the pixels through the initialized preset frog leaping algorithm to obtain the The local optimal pixel points of the original image;
- the image segmentation module is configured to determine each of the local optimal pixel points as the initial clustering center of the fuzzy C-means clustering algorithm, and use the fuzzy C-means clustering algorithm to compare the original The image is re-clustered to obtain segmented images.
- a third aspect of the embodiments of the present application provides a computer-readable storage medium, the computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the foregoing first aspect is implemented The steps of the image segmentation method.
- the fourth aspect of the embodiments of the present application provides a terminal device, including a memory, a processor, and computer-readable instructions stored in the memory and running on the processor, and the processor executes the The following steps are implemented when computer-readable instructions:
- the preset leaping algorithm is initialized according to the number of clusters and the pixel points, and the pixel points are optimized through the preset leaping algorithm after initialization to obtain the local optimum of the original image pixel;
- the preset frog leaping algorithm can be initialized first according to the number of clusters and pixels, and then the preset frog after initialization can be used. Jump algorithm to optimize the pixels, get each local optimal pixel of the original image, and determine the obtained local optimal pixel as the initial cluster center of the fuzzy C-means clustering algorithm, and then use the fuzzy C-means
- the clustering algorithm performs image segmentation on the original image according to the initial cluster center.
- the initial clustering center of the fuzzy C-means clustering algorithm can be determined by combining the preset frog-leapfrog algorithm, and the initial clustering center is used for re-clustering to realize image segmentation, and the fuzzy C can be reasonably determined The initial clustering center of the mean clustering algorithm to improve the image segmentation effect.
- the frog leaping algorithm has the characteristics of fast calculation speed and strong optimization ability, so the preset frog leaping algorithm is combined with the fuzzy C-means clustering algorithm To perform image segmentation, it can effectively reduce the computational complexity in the image segmentation process, increase the convergence speed, and greatly improve the image segmentation efficiency and segmentation effect.
- FIG. 1 is a flowchart of an embodiment of an image segmentation method in an embodiment of the application
- FIG. 2 is a schematic diagram of a flow chart of an image segmentation method in an application scenario for optimization processing in an embodiment of this application;
- FIG. 3 is a structural diagram of an embodiment of an image segmentation device in an embodiment of the application.
- Fig. 4 is a schematic diagram of a terminal device provided by an embodiment of the application.
- the embodiments of the application provide an image segmentation method, device, computer readable storage medium, and terminal equipment, which are used to reasonably determine the initial cluster center of FCM clustering, improve image segmentation efficiency and segmentation effect, and can effectively reduce image segmentation The computational complexity of the process.
- an embodiment of the present application provides an image segmentation method, and the image segmentation method includes:
- Step S101 Obtain an original image to be segmented, and determine the number of clusters and pixels of the original image;
- the execution subject of the embodiments of the present application is a terminal device, which includes but is not limited to: servers, computers, smart phones, and tablet computers.
- the original image to be segmented can be input into the terminal device, and the number of clusters of the original image can be set.
- the number of clusters of the original image can be set to 4 or 3, etc.
- the terminal device can determine the pixels of the original image, such as determining the number of pixels and the position of each pixel of the original image.
- Step S102 Initialize a preset leapfrog algorithm according to the number of clusters and the pixels, and perform optimization processing on the pixels through the preset leapfrog algorithm after initialization to obtain each of the original images Local optimal pixel points;
- the terminal device can initialize a preset frog leaping algorithm according to the number of clusters and the pixels, and pass the initialization
- the preset frog-leapfrog algorithm performs optimization processing on the pixel points to obtain each local optimal pixel point of the original image.
- the preset leaping algorithm is initialized according to the number of clusters and the pixel points, and the pixel points are optimized by the preset leaping algorithm after initialization.
- Obtaining the local optimal pixel points of the original image may include:
- Step S201 Determine the number of clusters as the number of population groups of the preset leapfrog algorithm, construct an initial population of the number of population groups, and initialize the first cluster center of each of the initial populations;
- the number of clusters of the original image can be determined as the number of population groups of the preset frog leaping algorithm to construct the initial population of the number of population groups, and the first cluster of each initial population can be initialized.
- the cluster center for example, the first cluster center of each initial population can be generated by the random generation function rand(i).
- the number of clusters of the original image is 4, then the The number of population groups of the preset frog leaping algorithm is 4, which means that 4 initial populations need to be constructed, that is, 4 first cluster centers need to be generated by the random generation function rand(i), and each first cluster The center corresponds to an initial population.
- Step S202 Determine the frog individuals of the preset frog leaping algorithm according to the pixels, and calculate the first degree of membership between each frog individual and each of the first cluster centers;
- a preset number of frog individuals can be generated according to the pixels of the original image.
- NI pixels can be selected from the pixels of the original image as the frog individuals in the preset leaping algorithm. That is to say, each selected pixel is regarded as an individual frog, wherein the preset number NI can be determined according to the initialization value of the preset frog leaping algorithm, that is to say, all can be performed in the terminal device in advance.
- the initialization of the preset leapfrog algorithm for example, the group size NI of the preset leapfrog algorithm, the maximum number of iterations J in the group, the maximum variable Dmax allowed when the frog individual changes position, etc. can be preset.
- the membership degree between each frog individual and each first clustering center can be calculated.
- the membership degree calculation of the fuzzy C-means clustering algorithm can be used
- the formula calculates the first degree of membership between each frog individual and each of the first cluster centers. Specifically, the formula for calculating the degree of membership is:
- u ij between the first membership frog individual first cluster center j and i X j is the position vector of the j individual frog
- V i is the position vector of the first cluster center i
- C is the cluster number
- m is the fuzzy blur degree C means clustering algorithm
- is the Euclidean distance between X j and V i.
- the ambiguity m of the fuzzy C-means clustering algorithm may also be set in the terminal device in advance.
- Step S203 dividing each frog individual into a corresponding initial population according to the first degree of membership, and obtaining the worst frog individual of each of the initial population according to the first degree of membership;
- each frog individual can be divided into the corresponding initial population according to each first degree of membership.
- Maximum membership degree and determine the initial population of the first cluster center corresponding to each maximum membership degree as the initial population of each frog individual, that is, divide each frog individual into the first cluster center corresponding to its maximum membership In the initial population.
- the first cluster center includes A, B, C, and D.
- the first degree of membership between the first frog individual and the first cluster center A is 0.6
- the The first degree of membership between a cluster center B is 0.3
- the first degree of membership with the first cluster center C is 0.4
- the first degree of membership with the first cluster center D is 0.1
- the first membership degree between the second frog individual and the first cluster center A is 0.4
- the first membership degree between the second frog individual and the first cluster center B is 0.1
- the first membership degree between the second frog individual and the first cluster center C When the degree of membership is 0.5 and the first degree of membership with the first cluster center D is 0, the first individual frog can be divided into the initial population A to which the first cluster center A belongs, and the second Each frog individual is divided into the initial population C to which the first cluster center C belongs.
- the worst frog individual in each initial population can be obtained, where the worst frog individual is the frog individual with the first minimum degree of membership in each initial population.
- Step S204 Perform a position update on each of the worst frog individuals according to a preset update method to obtain an updated new population
- the position of each worst frog individual can be updated.
- the position update method based on the shrinkage factor ⁇ can be used to update the position of each worst frog individual. , Get the updated new population.
- the location update method based on the shrinkage factor ⁇ may specifically be:
- newX i is the position vector of the individual i update frog
- ⁇ is the shrinkage factor
- X i is the position vector before updating frog individual i
- D is the frog individual coefficient update step size.
- the update step coefficient can be specifically set according to actual conditions.
- the new membership degree between the updated worst frog individual and the corresponding first cluster center can be calculated. And determine whether the new membership degree satisfies a preset condition, such as determining whether the new membership degree is greater than the first membership degree of the worst frog individual before the location update, if the new membership degree satisfies the preset condition, If the new membership degree is greater than the first membership degree of the worst frog individual before the location update, the worst frog individual after the location update is retained in the initial population, and the retained initial population is determined as the updated If the new membership degree does not meet the preset conditions, if the new membership degree is less than or equal to the first membership degree of the worst frog individual before the location update, a new frog can be randomly generated Individual, and replace the worst frog individual with the new frog individual to obtain an updated new population.
- a preset condition such as determining whether the new membership degree is greater than the first membership degree of the worst frog individual before the location update, if the new membership degree sati
- Step S205 Determine whether the new population meets a preset termination condition
- the preset termination condition may be whether the number of iterations of each group reaches a preset group
- the maximum number of iterations J can also be whether the optimal frog individuals in each group have the same algebraic number continuously reaching the preset algebraic value, or whether the objective function value of each new population is less than a preset preset threshold, and so on.
- the preset termination condition is preferably whether the objective function value is less than a preset threshold. Therefore, the judging whether the new population meets the preset termination condition may include:
- F(t) is the objective function value
- C is the number of clusters
- N is the total number of frog individuals
- u ij is the first degree of membership between the frog individual j and the first cluster center i
- X j is frog of individual position vectors j
- V i is a position vector i in a first cluster center.
- the objective function value corresponding to the new population can be obtained according to the above-mentioned objective function value calculation formula, and it can be judged whether the objective function value meets the preset condition, such as judging whether the objective function value is less than the preset condition Threshold to determine whether the new population meets the preset termination conditions.
- Step S206 If the new population satisfies the preset termination condition, obtain the optimal frog individuals in each of the new populations, and determine each of the optimal frog individuals as the local optimal of the original image pixel;
- Step S207 If the new population does not meet the preset termination condition, determine the second cluster center of each new population, and calculate each frog individual in each new population and the corresponding second cluster The second degree of membership between the centers, and each of the new populations is determined as the initial population, the second degree of membership is determined as the first degree of membership, and the execution returns to obtaining each of the initial populations according to the first degree of membership The worst frog individual steps and the next steps.
- the updated new population meets the preset termination condition, for example, the number of iterations within the group of each new population reaches the preset maximum number of iterations, or the new population corresponds to If the objective function value of is less than the preset threshold, you can terminate the iteration and obtain the optimal frog individuals of each new population at this time, that is, obtain the first frog individuals with the largest membership degree in each new population, and combine the obtained maximum
- the pixels represented by the excellent frog individuals are determined as the local optimal pixels of the original image; if the new population does not meet the preset termination conditions, if the number of iterations within the group of a new population is less than the preset maximum number of iterations , Or if the objective function value corresponding to the new population is greater than or equal to the preset threshold, the second clustering center of each new population can be re-determined, and the second membership degree of each frog individual can be recalculated in each new population,
- the determining the second cluster center of each of the new populations may include:
- V s is the second clustering center of the s-th new population
- u ts is the membership degree between frog individuals t and V s in the s-th new population
- m is the ambiguity of the fuzzy C-means clustering algorithm
- X t is the position vector of frog individual t
- T is the total number of frog individuals in the s-th new population.
- Step S103 Determine each of the local optimal pixels as the initial clustering center of the fuzzy C-means clustering algorithm, and use the fuzzy C-means clustering algorithm to reconstruct the original image according to the initial clustering center. Clustering to obtain segmented images.
- each local optimal pixel point of the original image can be determined as the initial cluster center of the fuzzy C-means clustering algorithm .
- the fuzzy C-means clustering algorithm to re-cluster the original image according to the initial clustering center to obtain a segmented image, wherein the number of acquired local optimal pixels is the number of clusters , For example, when the number of clusters is 4, 4 local optimal pixels can be obtained.
- the process of the fuzzy C-means clustering algorithm to re-cluster the original image according to the initial clustering center for image segmentation is the same as that of the traditional fuzzy C-means clustering algorithm for image segmentation.
- the method of determining the initial clustering center in the traditional fuzzy C-means clustering algorithm is mainly optimized through the preset frog-leapfrog algorithm, so as to improve the accuracy of determining the initial clustering center and improve the image segmentation effect and efficiency. .
- the preset frog leaping algorithm can be initialized first according to the number of clusters and pixels, and then the preset frog after initialization can be used. Jump algorithm to optimize the pixels, get each local optimal pixel of the original image, and determine the obtained local optimal pixel as the initial cluster center of the fuzzy C-means clustering algorithm, and then use the fuzzy C-means
- the clustering algorithm performs image segmentation on the original image according to the initial cluster center.
- the initial clustering center of the fuzzy C-means clustering algorithm can be determined by combining the preset frog-leapfrog algorithm, and the initial clustering center is used for re-clustering to realize image segmentation, and the fuzzy C can be reasonably determined The initial clustering center of the mean clustering algorithm to improve the image segmentation effect.
- the frog leaping algorithm has the characteristics of fast calculation speed and strong optimization ability, so the preset frog leaping algorithm is combined with the fuzzy C-means clustering algorithm To perform image segmentation, it can effectively reduce the computational complexity in the image segmentation process, increase the convergence speed, and greatly improve the image segmentation efficiency and segmentation effect.
- an embodiment of the present application provides an image segmentation device, and the image segmentation device includes:
- the original image obtaining module 301 is used to obtain the original image to be segmented and determine the number of clusters and pixels of the original image;
- the optimization processing module 302 is configured to initialize a preset frog leaping algorithm according to the number of clusters and the pixels, and perform optimization processing on the pixels through the initialized preset frog leaping algorithm to obtain all State the local optimal pixel points of the original image;
- the image segmentation module 303 is configured to determine each of the local optimal pixels as the initial clustering center of the fuzzy C-means clustering algorithm, and use the fuzzy C-means clustering algorithm to compare the The original image is re-clustered to obtain segmented images.
- optimization processing module 302 may include:
- the initial population construction unit is configured to determine the number of clusters as the number of population groups of the preset leaping algorithm, construct an initial population of the number of population groups, and initialize the first cluster of each of the initial populations Class center
- the first membership degree calculation unit is configured to determine the frog individuals of the preset leapfrog algorithm according to the pixel points, and calculate the first membership degree between each frog individual and each of the first cluster centers;
- An individual division unit configured to divide each frog individual into a corresponding initial population according to the first degree of membership, and obtain the worst frog individual of each of the initial population according to the first degree of membership;
- the new population obtaining unit is configured to update the position of each of the worst frog individuals according to a preset update method to obtain an updated new population
- a new population judging unit for judging whether the new population meets a preset termination condition
- the optimal pixel point determination unit is configured to, if the new population meets the preset termination condition, obtain the optimal frog individuals in each of the new populations, and determine each of the optimal frog individuals as the original The local optimal pixel points of the image;
- the second membership degree calculation unit is configured to determine the second cluster center of each new population if the new population does not meet the preset termination condition, and calculate the relationship between each frog individual in each new population Corresponding to the second degree of membership between the second cluster centers, and determine each of the new populations as the initial population, determine the second degree of membership as the first degree of membership, and return to execute according to the first degree of membership The steps of obtaining the worst frog individuals of each of the initial populations and subsequent steps.
- the first membership degree calculation unit is specifically configured to use the membership degree calculation formula of the fuzzy C-means clustering algorithm to calculate the first membership degree between each frog individual and each of the first cluster centers , wherein the calculation formula for the degree of membership is:
- u ij between the first membership frog individual first cluster center j and i X j is the position vector of the j individual frog
- V i is the position vector of the first cluster center i
- C is the cluster number
- m is the fuzzy blur degree C means clustering algorithm
- is the Euclidean distance between X j and V i.
- the new population judgment unit is specifically configured to calculate an objective function value of the new population according to each of the first membership degrees, and judge whether the objective function value meets a preset condition;
- F(t) is the objective function value
- C is the number of clusters
- N is the total number of frog individuals
- u ij is the first degree of membership between the frog individual j and the first cluster center i
- X j is frog of individual position vectors j
- V i is a position vector i in a first cluster center.
- the new population acquisition unit is specifically configured to update the position of each worst frog individual according to the following update formula:
- D is the coefficient update step size frog subject.
- the second membership degree calculation unit is specifically configured to determine the second cluster center of each new population according to the following determination formula:
- V s is the second clustering center of the s-th new population
- u ts is the membership degree between frog individuals t and V s in the s-th new population
- m is the ambiguity of the fuzzy C-means clustering algorithm
- X t is the position vector of frog individual t
- T is the total number of frog individuals in the s-th new population.
- Fig. 4 is a schematic diagram of a terminal device provided by an embodiment of the present application.
- the terminal device 4 of this embodiment includes: a processor 40, a memory 41, and computer-readable instructions 42 stored in the memory 41 and running on the processor 40, such as an image segmentation program .
- the processor 40 executes the computer-readable instructions 42, the steps in the foregoing image segmentation method embodiments are implemented, such as steps S101 to S103 shown in FIG. 1.
- the processor 40 executes the computer-readable instructions 42
- the functions of the modules/units in the foregoing device embodiments such as the functions of the modules 301 to 303 shown in FIG. 3, are realized.
- the computer-readable instructions 42 may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 41 and executed by the processor 40, To complete this application.
- the one or more modules/units may be a series of computer-readable instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer-readable instructions 42 in the terminal device 4.
- the terminal device 4 may be a computing device such as a desktop computer, a notebook, a palmtop computer, and a cloud server.
- the terminal device may include, but is not limited to, a processor 40 and a memory 41.
- FIG. 4 is only an example of the terminal device 4, and does not constitute a limitation on the terminal device 4. It may include more or less components than shown in the figure, or a combination of certain components, or different components.
- the terminal device may also include input and output devices, network access devices, buses, etc.
- the processor 40 may be a central processing unit (Central Processing Unit, CPU), other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (ASIC), Ready-made programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
- the general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like.
- the memory 41 may be an internal storage unit of the terminal device 4, such as a hard disk or memory of the terminal device 4.
- the memory 41 may also be an external storage device of the terminal device 4, such as a plug-in hard disk equipped on the terminal device 4, a smart memory card (Smart Media Card, SMC), and a Secure Digital (SD) Card, Flash Card, etc. Further, the memory 41 may also include both an internal storage unit of the terminal device 4 and an external storage device.
- the memory 41 is used to store the computer-readable instructions and other programs and data required by the terminal device.
- the memory 41 can also be used to temporarily store data that has been output or will be output.
- each unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit.
- the above-mentioned integrated unit can be implemented in the form of hardware or software functional unit.
- Non-volatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory.
- Volatile memory may include random access memory (RAM) or external cache memory.
- RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Channel (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
Description
Claims (20)
- 一种图像分割方法,其特征在于,包括:获取待分割的原始图像,并确定所述原始图像的聚类数目和像素点;根据所述聚类数目和所述像素点对预设蛙跳算法进行初始化,并通过初始化后的预设蛙跳算法对所述像素点进行寻优处理,得到所述原始图像的各局部最优像素点;将各所述局部最优像素点确定为模糊C均值聚类算法的初始聚类中心,并利用所述模糊C均值聚类算法根据所述初始聚类中心对所述原始图像进行再聚类,得到分割图像。
- 根据权利要求1所述的图像分割方法,其特征在于,所述根据所述聚类数目和所述像素点对预设蛙跳算法进行初始化,并通过初始化后的预设蛙跳算法对所述像素点进行寻优处理,得到所述原始图像的各局部最优像素点,包括:将所述聚类数目确定为所述预设蛙跳算法的种群组数,构建所述种群组数的初始种群,并初始化各所述初始种群的第一聚类中心;根据所述像素点确定所述预设蛙跳算法的青蛙个体,并计算每一个青蛙个体与各所述第一聚类中心之间的第一隶属度;根据所述第一隶属度将各青蛙个体划分至对应的初始种群,并根据所述第一隶属度获取各所述初始种群的最差青蛙个体;按照预设更新方式对各所述最差青蛙个体进行位置更新,得到更新后的新种群;判断所述新种群是否满足预设终止条件;若所述新种群满足所述预设终止条件,则获取各所述新种群中的最优青蛙个体,并将各所述最优青蛙个体确定为所述原始图像的各局部最优像素点;若所述新种群不满足所述预设终止条件,则确定各所述新种群的第二聚类中心,计算各所述新种群中的每一个青蛙个体与对应的第二聚类中心之间的第二隶属度,并将各所述新种群确定为初始种群、将所述第二隶属度确定为第一隶属度,返回执行根据所述第一隶属度获取各所述初始种群的最差青蛙个体的步骤以及后续步骤。
- 一种图像分割装置,其特征在于,包括:原始图像获取模块,用于获取待分割的原始图像,并确定所述原始图像的聚类数目和像素点;寻优处理模块,用于根据所述聚类数目和所述像素点对预设蛙跳算法进行初始化,并通过初始化后的预设蛙跳算法对所述像素点进行寻优处理,得到所述原始图像的各局部最优像素点;图像分割模块,用于将各所述局部最优像素点确定为模糊C均值聚类算法的初始聚类中心,并利用所述模糊C均值聚类算法根据所述初始聚类中心对所述原始图像进行再聚类,得到分割图像。
- 根据权利要求7所述的图像分割装置,其特征在于,所述寻优处理模块,包括:初始种群构建单元,用于将所述聚类数目确定为所述预设蛙跳算法的种群组数,构建所述种群组数的初始种群,并初始化各所述初始种群的第一聚类中心;第一隶属度计算单元,用于根据所述像素点确定所述预设蛙跳算法的青蛙个体,并计算每一个青蛙个体与各所述第一聚类中心之间的第一隶属度;个体划分单元,用于根据所述第一隶属度将各青蛙个体划分至对应的初始种群,并根据所述第一隶属度获取各所述初始种群的最差青蛙个体;新种群获取单元,用于按照预设更新方式对各所述最差青蛙个体进行位置更新,得到更新后的新种群;新种群判断单元,用于判断所述新种群是否满足预设终止条件;最优像素点确定单元,用于若所述新种群满足所述预设终止条件,则获取各所述新种群中的最优青蛙个体,并将各所述最优青蛙个体确定为所述原始图像的各局部最优像素点;第二隶属度计算单元,用于若所述新种群不满足所述预设终止条件,则确定各所述新种群的第二聚类中心,计算各所述新种群中的每一个青蛙个体与对应的第二聚类中心之间的第二隶属度,并将各所述新种群确定为初始种群、将所述第二隶属度确定为第一隶属度,返回执行根据所述第一隶属度获取各所述初始种群的最差青蛙个体的步骤以及后续步骤。
- 一种计算机可读存储介质,所述计算机可读存储介质存储有计算机可读指令,其特征在于,所述计算机可读指令被处理器执行时实现如下步骤:获取待分割的原始图像,并确定所述原始图像的聚类数目和像素点;根据所述聚类数目和所述像素点对预设蛙跳算法进行初始化,并通过初始化后的预设蛙跳算法对所述像素点进行寻优处理,得到所述原始图像的各局部最优像素点;将各所述局部最优像素点确定为模糊C均值聚类算法的初始聚类中心,并利用所述模糊C均值聚类算法根据所述初始聚类中心对所述原始图像进行再聚类,得到分割图像。
- 根据权利要求13所述的计算机可读存储介质,其特征在于,所述根据所述聚类数目和所述像素点对预设蛙跳算法进行初始化,并通过初始化后的预设蛙跳算法对所述像素点进行寻优处理,得到所述原始图像的各局部最优像素点,包括:将所述聚类数目确定为所述预设蛙跳算法的种群组数,构建所述种群组数的初始种群,并初始化各所述初始种群的第一聚类中心;根据所述像素点确定所述预设蛙跳算法的青蛙个体,并计算每一个青蛙个体与各所述第一聚类中心之间的第一隶属度;根据所述第一隶属度将各青蛙个体划分至对应的初始种群,并根据所述第一隶属度获取各所述初始种群的最差青蛙个体;按照预设更新方式对各所述最差青蛙个体进行位置更新,得到更新后的新种群;判断所述新种群是否满足预设终止条件;若所述新种群满足所述预设终止条件,则获取各所述新种群中的最优青蛙个体,并将各所述最优青蛙个体确定为所述原始图像的各局部最优像素点;若所述新种群不满足所述预设终止条件,则确定各所述新种群的第二聚类中心,计算各所述新种群中的每一个青蛙个体与对应的第二聚类中心之间的第二隶属度,并将各所述新种群确定为初始种群、将所述第二隶属度确定为第一隶属度,返回执行根据所述第一隶属度获取各所述初始种群的最差青蛙个体的步骤以及后续步骤。
- 一种终端设备,包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令,其特征在于,所述处理器执行所述计算机可读指令时实现如下步骤:获取待分割的原始图像,并确定所述原始图像的聚类数目和像素点;根据所述聚类数目和所述像素点对预设蛙跳算法进行初始化,并通过初始化后的预设蛙跳算法对所述像素点进行寻优处理,得到所述原始图像的各局部最优像素点;将各所述局部最优像素点确定为模糊C均值聚类算法的初始聚类中心,并利用所述模糊C均值聚类算法根据所述初始聚类中心对所述原始图像进行再聚类,得到分割图像。
- 根据权利要求15所述的终端设备,其特征在于,所述根据所述聚类数目和所述像素点对预设蛙跳算法进行初始化,并通过初始化后的预设蛙跳算法对所述像素点进行寻优处理,得到所述原始图像的各局部最优像素点,包括:将所述聚类数目确定为所述预设蛙跳算法的种群组数,构建所述种群组数的初始种群,并初始化各所述初始种群的第一聚类中心;根据所述像素点确定所述预设蛙跳算法的青蛙个体,并计算每一个青蛙个体与各所述第一聚类中心之间的第一隶属度;根据所述第一隶属度将各青蛙个体划分至对应的初始种群,并根据所述第一隶属度获取各所述初始种群的最差青蛙个体;按照预设更新方式对各所述最差青蛙个体进行位置更新,得到更新后的新种群;判断所述新种群是否满足预设终止条件;若所述新种群满足所述预设终止条件,则获取各所述新种群中的最优青蛙个体,并将各所述最优青蛙个体确定为所述原始图像的各局部最优像素点;若所述新种群不满足所述预设终止条件,则确定各所述新种群的第二聚类中心,计算各所述新种群中的每一个青蛙个体与对应的第二聚类中心之间的第二隶属度,并将各所述新种群确定为初始种群、将所述第二隶属度确定为第一隶属度,返回执行根据所述第一隶属度获取各所述初始种群的最差青蛙个体的步骤以及后续步骤。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910422757.2 | 2019-05-21 | ||
CN201910422757.2A CN110264471A (zh) | 2019-05-21 | 2019-05-21 | 一种图像分割方法、装置、存储介质及终端设备 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2020233084A1 true WO2020233084A1 (zh) | 2020-11-26 |
Family
ID=67914904
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2019/122572 WO2020233084A1 (zh) | 2019-05-21 | 2019-12-03 | 一种图像分割方法、装置、存储介质及终端设备 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN110264471A (zh) |
WO (1) | WO2020233084A1 (zh) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112508963A (zh) * | 2020-11-30 | 2021-03-16 | 西安电子科技大学 | 一种基于模糊c均值聚类的sar图像分割方法 |
CN113705658A (zh) * | 2021-08-24 | 2021-11-26 | 浙江中烟工业有限责任公司 | 一种烟丝杂物智能识别方法 |
CN113901995A (zh) * | 2021-09-28 | 2022-01-07 | 广东奥博信息产业股份有限公司 | 一种面向环境噪声数据的聚类方法及系统 |
CN114330526A (zh) * | 2021-12-23 | 2022-04-12 | 北京交通大学 | 基于tgo-ssa-fcm算法的交通流量缺失数据修复方法 |
CN114596438A (zh) * | 2022-03-10 | 2022-06-07 | 南京邮电大学 | 一种求解基因芯片图像分割的Memetic方法 |
CN114595422A (zh) * | 2022-02-24 | 2022-06-07 | 大族激光科技产业集团股份有限公司 | 一种加工头蛙跳计算方法、加工设备及存储介质 |
CN115952432A (zh) * | 2022-12-21 | 2023-04-11 | 四川大学华西医院 | 一种基于糖尿病数据的无监督聚类方法 |
CN116188496A (zh) * | 2023-04-25 | 2023-05-30 | 牧马人(山东)勘察测绘集团有限公司 | 基于土地利用类型的遥感影像自适应分割方法 |
CN116260547A (zh) * | 2023-05-11 | 2023-06-13 | 武汉能钠智能装备技术股份有限公司四川省成都市分公司 | 一种抑制同频干扰的系统及方法 |
CN116309575A (zh) * | 2023-05-19 | 2023-06-23 | 济宁众达利电气设备有限公司 | 基于图像处理的电插头生产质量检测方法 |
CN117408394A (zh) * | 2023-12-14 | 2024-01-16 | 国网天津市电力公司电力科学研究院 | 电力系统的碳排放因子预测方法、装置及电子设备 |
CN117828958A (zh) * | 2024-03-06 | 2024-04-05 | 深圳大学 | 基于蟋蟀争斗优化算法的桁架结构优化设计方法和装置 |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110969170B (zh) * | 2019-12-03 | 2024-03-08 | 北京奇艺世纪科技有限公司 | 一种图像主题色提取方法、装置及电子设备 |
CN116563326A (zh) * | 2023-04-28 | 2023-08-08 | 江西绿萌科技控股有限公司 | 一种面向果蔬分选设备的多目标图像分割方法 |
CN117115754B (zh) * | 2023-10-23 | 2023-12-26 | 贵州省种畜禽种质测定中心 | 基于计算机视觉的鸭舍智能化监测方法 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101599176A (zh) * | 2009-06-17 | 2009-12-09 | 中国科学院自动化研究所 | 一种对管状结构内层分割的方法 |
CN102800089A (zh) * | 2012-06-28 | 2012-11-28 | 华中科技大学 | 基于颈部超声图像的主颈动脉血管提取和厚度测量方法 |
CN102855633A (zh) * | 2012-09-05 | 2013-01-02 | 山东大学 | 一种具有抗噪性的快速模糊聚类数字图像分割方法 |
CN109447997A (zh) * | 2018-09-01 | 2019-03-08 | 哈尔滨工程大学 | 一种自适应文化算法的水下声纳图像目标分割方法 |
US10235750B2 (en) * | 2014-02-27 | 2019-03-19 | Agency For Science, Technology And Research | Segmentation of cardiac magnetic resonance (CMR) images using a memory persistence approach |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB9626676D0 (en) * | 1996-12-23 | 1997-02-12 | Smith & Nephew Res | Imaging method |
CN105931236B (zh) * | 2016-04-19 | 2018-12-14 | 武汉大学 | 面向图像分割的模糊c-均值聚类初始聚类中心自动选取方法 |
CN108805893A (zh) * | 2018-06-06 | 2018-11-13 | 甘肃农业大学 | 苹果图像分割方法及装置 |
-
2019
- 2019-05-21 CN CN201910422757.2A patent/CN110264471A/zh active Pending
- 2019-12-03 WO PCT/CN2019/122572 patent/WO2020233084A1/zh active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101599176A (zh) * | 2009-06-17 | 2009-12-09 | 中国科学院自动化研究所 | 一种对管状结构内层分割的方法 |
CN102800089A (zh) * | 2012-06-28 | 2012-11-28 | 华中科技大学 | 基于颈部超声图像的主颈动脉血管提取和厚度测量方法 |
CN102855633A (zh) * | 2012-09-05 | 2013-01-02 | 山东大学 | 一种具有抗噪性的快速模糊聚类数字图像分割方法 |
US10235750B2 (en) * | 2014-02-27 | 2019-03-19 | Agency For Science, Technology And Research | Segmentation of cardiac magnetic resonance (CMR) images using a memory persistence approach |
CN109447997A (zh) * | 2018-09-01 | 2019-03-08 | 哈尔滨工程大学 | 一种自适应文化算法的水下声纳图像目标分割方法 |
Non-Patent Citations (3)
Title |
---|
DONGLING LIU ET AL.: ""Medical Image Segmentation based on Improved Fuzzy C-means Clustering"", 2017 INTERNATIONAL CONFERENCE ON SMART GRID AND ELECTRICAL AUTOMATION (ICSGEA), 28 May 2017 (2017-05-28), pages 406 - 410, XP033253339 * |
GU YING-JIE; JIA ZHEN-HONG; QIN XI-ZHONG; YANG JIE; PANG SHAO-NING: "Image segmentation algorithm based on shuffled frog-leaping with FCM", COMMUNICATIONS TECHNOLOGY, vol. 44, no. 2, 28 February 2011 (2011-02-28), pages 118 - 119+122, XP009524357, ISSN: 1002-0802 * |
LI, ZHEN , LUO KE : "Improved FCM algorithm based on PSO and SFLA", JOURNAL OF COMPUTER APPLICATIONS, vol. 31, no. 5, 31 May 2011 (2011-05-31), pages 1355 - 1358, XP009524361, ISSN: 1001-9081, DOI: 10.3724/SP.J.1087.2011.01355 * |
Cited By (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112508963B (zh) * | 2020-11-30 | 2023-09-19 | 西安电子科技大学 | 一种基于模糊c均值聚类的sar图像分割方法 |
CN112508963A (zh) * | 2020-11-30 | 2021-03-16 | 西安电子科技大学 | 一种基于模糊c均值聚类的sar图像分割方法 |
CN113705658A (zh) * | 2021-08-24 | 2021-11-26 | 浙江中烟工业有限责任公司 | 一种烟丝杂物智能识别方法 |
CN113705658B (zh) * | 2021-08-24 | 2024-03-19 | 浙江中烟工业有限责任公司 | 一种烟丝杂物智能识别方法 |
CN113901995A (zh) * | 2021-09-28 | 2022-01-07 | 广东奥博信息产业股份有限公司 | 一种面向环境噪声数据的聚类方法及系统 |
CN113901995B (zh) * | 2021-09-28 | 2024-04-12 | 广东奥博信息产业股份有限公司 | 一种面向环境噪声数据的聚类方法及系统 |
CN114330526A (zh) * | 2021-12-23 | 2022-04-12 | 北京交通大学 | 基于tgo-ssa-fcm算法的交通流量缺失数据修复方法 |
CN114330526B (zh) * | 2021-12-23 | 2024-04-19 | 北京交通大学 | 基于tgo-ssa-fcm算法的交通流量缺失数据修复方法 |
CN114595422A (zh) * | 2022-02-24 | 2022-06-07 | 大族激光科技产业集团股份有限公司 | 一种加工头蛙跳计算方法、加工设备及存储介质 |
CN114595422B (zh) * | 2022-02-24 | 2024-05-31 | 大族激光科技产业集团股份有限公司 | 一种加工头蛙跳计算方法、加工设备及存储介质 |
CN114596438A (zh) * | 2022-03-10 | 2022-06-07 | 南京邮电大学 | 一种求解基因芯片图像分割的Memetic方法 |
CN115952432B (zh) * | 2022-12-21 | 2024-03-12 | 四川大学华西医院 | 一种基于糖尿病数据的无监督聚类方法 |
CN115952432A (zh) * | 2022-12-21 | 2023-04-11 | 四川大学华西医院 | 一种基于糖尿病数据的无监督聚类方法 |
CN116188496A (zh) * | 2023-04-25 | 2023-05-30 | 牧马人(山东)勘察测绘集团有限公司 | 基于土地利用类型的遥感影像自适应分割方法 |
CN116188496B (zh) * | 2023-04-25 | 2023-07-07 | 牧马人(山东)勘察测绘集团有限公司 | 基于土地利用类型的遥感影像自适应分割方法 |
CN116260547B (zh) * | 2023-05-11 | 2023-09-22 | 武汉能钠智能装备技术股份有限公司四川省成都市分公司 | 一种抑制同频干扰的系统及方法 |
CN116260547A (zh) * | 2023-05-11 | 2023-06-13 | 武汉能钠智能装备技术股份有限公司四川省成都市分公司 | 一种抑制同频干扰的系统及方法 |
CN116309575A (zh) * | 2023-05-19 | 2023-06-23 | 济宁众达利电气设备有限公司 | 基于图像处理的电插头生产质量检测方法 |
CN117408394A (zh) * | 2023-12-14 | 2024-01-16 | 国网天津市电力公司电力科学研究院 | 电力系统的碳排放因子预测方法、装置及电子设备 |
CN117408394B (zh) * | 2023-12-14 | 2024-05-31 | 国网天津市电力公司电力科学研究院 | 电力系统的碳排放因子预测方法、装置及电子设备 |
CN117828958A (zh) * | 2024-03-06 | 2024-04-05 | 深圳大学 | 基于蟋蟀争斗优化算法的桁架结构优化设计方法和装置 |
CN117828958B (zh) * | 2024-03-06 | 2024-05-17 | 深圳大学 | 基于蟋蟀争斗优化算法的桁架结构优化设计方法和装置 |
Also Published As
Publication number | Publication date |
---|---|
CN110264471A (zh) | 2019-09-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2020233084A1 (zh) | 一种图像分割方法、装置、存储介质及终端设备 | |
WO2023000574A1 (zh) | 一种模型训练方法、装置、设备及可读存储介质 | |
CN112580416A (zh) | 基于深暹罗网络和贝叶斯优化的视频跟踪 | |
CN112639843A (zh) | 使用机器学习模型来抑制偏差数据 | |
WO2021089013A1 (zh) | 空间图卷积网络的训练方法、电子设备及存储介质 | |
CN111553215B (zh) | 人员关联方法及其装置、图卷积网络训练方法及其装置 | |
CN112037912A (zh) | 基于医疗知识图谱的分诊模型训练方法、装置及设备 | |
WO2022042123A1 (zh) | 图像识别模型生成方法、装置、计算机设备和存储介质 | |
WO2016062044A1 (zh) | 一种模型参数训练方法、装置及系统 | |
WO2021164269A1 (zh) | 基于注意力机制的视差图获取方法和装置 | |
EP3839764A1 (en) | Method and system for detecting duplicate document using vector quantization | |
WO2023020113A1 (zh) | 一种检测用户特征的方法、装置、设备及存储介质 | |
CN114677565B (zh) | 特征提取网络的训练方法和图像处理方法、装置 | |
CN111383232A (zh) | 抠图方法、装置、终端设备及计算机可读存储介质 | |
CN110246095B (zh) | 一种图像修复方法、装置、存储介质及终端设备 | |
CN111259137A (zh) | 知识图谱摘要的生成方法及系统 | |
WO2017088587A1 (zh) | 一种数据处理方法及装置 | |
WO2022127037A1 (zh) | 一种数据分类方法、装置及相关设备 | |
CN112036316B (zh) | 手指静脉识别方法、装置、电子设备及可读存储介质 | |
WO2021189845A1 (zh) | 时间序列异常点的检测方法、装置、设备及可读存储介质 | |
JP7207530B2 (ja) | 情報処理装置、作成方法および作成プログラム | |
Zhang et al. | An adaptive mean shift clustering algorithm based on locality-sensitive hashing | |
WO2021027162A1 (zh) | 一种非满格表格内容提取方法、装置及终端设备 | |
CN113515519A (zh) | 图结构估计模型的训练方法、装置、设备及存储介质 | |
WO2022257433A1 (zh) | 图像的特征图的处理方法及装置、存储介质、终端 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 19929617 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 19929617 Country of ref document: EP Kind code of ref document: A1 |
|
32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 31.03.2022) |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 19929617 Country of ref document: EP Kind code of ref document: A1 |