CN110866527A - Image segmentation method and device, electronic equipment and readable storage medium - Google Patents
Image segmentation method and device, electronic equipment and readable storage medium Download PDFInfo
- Publication number
- CN110866527A CN110866527A CN201811632456.4A CN201811632456A CN110866527A CN 110866527 A CN110866527 A CN 110866527A CN 201811632456 A CN201811632456 A CN 201811632456A CN 110866527 A CN110866527 A CN 110866527A
- Authority
- CN
- China
- Prior art keywords
- dimensional vector
- objective function
- weighted
- value
- calculating
- 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.)
- Pending
Links
Images
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Probability & Statistics with Applications (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention discloses an image segmentation method, an image segmentation device, electronic equipment and a readable storage medium, relates to the technical field of image processing, and can more accurately segment an image to obtain a better segmentation effect. The image segmentation method comprises the following steps: determining the number of clustering categories; calculating the self gray value and the neighborhood gray value of each pixel point in the image, and generating a two-dimensional vector set based on the self gray value and the neighborhood gray value of each pixel point; weighting each two-dimensional vector in the two-dimensional vector set to obtain a weighted two-dimensional vector; initializing a clustering center based on the weighted two-dimensional vector; calculating the Euclidean distance between the weighted two-dimensional vector and the clustering center; calculating fuzzy membership; updating the clustering center according to the fuzzy membership degree; calculating an objective function value based on the fuzzy membership degree and the Euclidean distance between the weighted two-dimensional vector and a clustering center; judging whether the difference value between the objective function value and the objective function value after the last iteration is smaller than an objective function change threshold value or not; and if the difference value between the objective function value and the objective function value after the last iteration is smaller than the objective function change threshold, carrying out image segmentation according to the maximum membership criterion. The apparatus, electronic device, and readable storage medium include modules for performing the methods. The invention is suitable for image segmentation.
Description
Technical Field
The present invention relates to the field of image processing, and in particular, to an image segmentation method and apparatus, an electronic device, and a readable storage medium.
Background
Images are the main source of information acquired and exchanged by human beings, and therefore, the application field of image processing necessarily relates to aspects of human life and work. With the continuous expansion of the human activity range, the application field of image processing will be continuously expanded. The method is mainly applied to the aspects of agriculture and animal husbandry, forestry, environment, military, industry, medicine and the like.
With the development of computers and the development of mathematical theories, particularly the creation and improvement of discrete mathematical theories, technologies for processing images, i.e., image processing technologies, have been developed and rapidly developed. The image segmentation is a key step for image processing, and the existing fuzzy C-means clustering algorithm (FCM) for segmenting the image can automatically realize clustering without manual operation, and belongs to an unsupervised mode in a classification method. However, when the fuzzy C-means clustering algorithm is used for segmenting and analyzing an image, only gray information of the image is considered, spatial information around pixels of the image is ignored, and meanwhile, when the distance between a sample point and a clustering center point is calculated, a Euclidean distance is used as a measurement reference, and the distance is sensitive to noise, burrs and discrete points, so that the segmentation result of the image by using the fuzzy C-means clustering algorithm is not accurate enough, and the segmentation effect is poor.
Disclosure of Invention
In view of this, embodiments of the present invention provide an image segmentation method, an image segmentation apparatus, an electronic device, and a readable storage medium, which can segment an image more accurately.
In a first aspect, an embodiment of the present invention provides an image segmentation method, including: determining the number of clustering categories; calculating the self gray value and the neighborhood gray value of each pixel point in the image, and generating a two-dimensional vector set based on the self gray value and the neighborhood gray value of each pixel point; weighting each two-dimensional vector in the two-dimensional vector set to obtain a weighted two-dimensional vector; initializing a clustering center based on the weighted two-dimensional vector; calculating the Euclidean distance between the weighted two-dimensional vector and the clustering center; calculating fuzzy membership; updating the clustering center according to the fuzzy membership degree; calculating an objective function value based on the fuzzy membership degree and the Euclidean distance between the weighted two-dimensional vector and a clustering center; judging whether the difference value between the objective function value and the objective function value after the last iteration is smaller than an objective function change threshold value or not; and if the difference value between the objective function value and the objective function value after the last iteration is smaller than the objective function change threshold, carrying out image segmentation according to the maximum membership criterion.
According to a specific implementation manner of the embodiment of the present invention, after determining whether a difference between the objective function value and the objective function value after the last iteration is smaller than an objective function change threshold, the method further includes:
and if the difference value between the objective function value and the objective function value after the last iteration is greater than or equal to the objective function change threshold, calculating the Euclidean distance between the weighted two-dimensional vector and the updated cluster center.
According to a specific implementation manner of the embodiment of the present invention, after calculating the objective function value based on the fuzzy membership matrix and the weighted euclidean distance, the method further includes:
and judging whether the iteration times reach a maximum iteration time threshold value, and if not, calculating the Euclidean distance between the weighted two-dimensional vector and the updated clustering center.
According to a specific implementation manner of the embodiment of the invention, calculating the neighborhood gray value of each pixel point in the image comprises the following steps:
the neighborhood gray value g (m, n) of each pixel point is calculated according to the following formula:
wherein s represents the width of a neighborhood window of a pixel point h (m, n);
m and n are coordinates of pixel points in the image, and i and j are relative coordinates of the pixel points in the width of the neighborhood window and the pixel points with the coordinates (m and n). h (m + i, n + j) is a pixel value with coordinates (m + i, n + j).
According to a specific implementation manner of the embodiment of the present invention, the calculating the euclidean distance between the weighted two-dimensional vector and the clustering center includes:
calculating the Euclidean distance between the weighted two-dimensional vector and the clustering center according to the following formula:
wherein, Fm，nFor any vector, V, in said set of weighted two-dimensional vectorspThe cluster center of the p-th class.
According to a specific implementation manner of the embodiment of the present invention, the calculating the fuzzy membership includes:
calculating the fuzzy membership degree of the kth weighted two-dimensional vector belonging to the pth clustering class according to the following formula:
wherein (F)m，n)kIs the k-th weighted two-dimensional vector; p is the pth cluster category; vpCluster centers for the p-th category; and c is the total number of the cluster categories.
According to a specific implementation manner of the embodiment of the present invention, the updating the clustering center according to the fuzzy membership degree includes:
calculating the cluster center of the p-type class category according to the following formula:
wherein f ism，nIs the two-dimensional vector; (u)m，n)kRepresenting the fuzzy membership degree of the kth weighted two-dimensional vector to each cluster class; a is a fuzzy weighting index; and N is the number of the two-dimensional vectors. p represents the p-thCategory of category.
According to a specific implementation manner of the embodiment of the present invention, the calculating an objective function value based on the fuzzy membership and the euclidean distance between the weighted two-dimensional vector and the clustering center includes:
the objective function value is calculated according to the following formula:
wherein the content of the first and second substances,
(um，n)kfuzzy membership from the kth weighted two-dimensional vector to the pth clustering class; a is a fuzzy weighting index.
In a second aspect, an embodiment of the present invention provides an image segmentation apparatus, including: the system comprises a category number determining module, a vector set generating module, a weighting processing module, a clustering center initializing module, a Euclidean distance calculating first module, a fuzzy membership calculating module, a clustering center updating module, an objective function value calculating module, a first judging module and an image segmentation module, wherein the category number determining module is used for determining the clustering category number; the vector set generation module is used for calculating the self gray value and the neighborhood gray value of each pixel point in the image and generating a two-dimensional vector matrix based on the self gray value and the neighborhood gray value of each pixel point; the weighting processing module is used for weighting each two-dimensional vector in the two-dimensional vector set to obtain a weighted two-dimensional vector; the cluster center initialization module is used for initializing the cluster center based on the weighted two-dimensional vector; the Euclidean distance calculation first module is used for calculating the Euclidean distance between the weighted two-dimensional vector and the clustering center; the fuzzy membership calculation module is used for calculating fuzzy membership; the cluster center updating module is used for updating the cluster center according to the fuzzy membership degree; the objective function value calculation module is used for calculating an objective function value based on the fuzzy membership degree and the Euclidean distance between the weighted two-dimensional vector and a clustering center; the first judgment module is used for judging whether the difference value between the objective function value and the objective function value after the last iteration is smaller than a target function change threshold value or not; and the image segmentation module is used for segmenting the image according to the maximum membership degree if the difference value between the objective function value and the objective function value after the last iteration is smaller than the objective function change threshold value.
According to a specific implementation manner of the embodiment of the present invention, the apparatus further includes:
and the Euclidean distance calculation second module is used for calculating the Euclidean distance between the weighted two-dimensional vector and the updated clustering center if the difference value between the objective function value and the objective function value after the last iteration is greater than or equal to the objective function change threshold value.
According to a specific implementation manner of the embodiment of the present invention, the apparatus further includes:
and the second judgment module is used for judging whether the iteration times reach a maximum iteration time threshold value, and if not, calculating the Euclidean distance between the weighted two-dimensional vector and the updated clustering center.
According to a specific implementation manner of the embodiment of the present invention, the vector matrix generation module includes:
the neighborhood gray level calculating module is used for calculating the neighborhood gray level value g (m, n) of each pixel point according to the following formula:
wherein s represents the width of a neighborhood window of a pixel point h (m, n);
m and n are coordinates of pixel points in the image, and i and j are relative coordinates of the pixel points in the width of the neighborhood window and the pixel points with the coordinates (m and n). h (m + i, n + j) is a pixel value with coordinates (m + i, n + j).
According to a specific implementation manner of the embodiment of the present invention, the euclidean distance calculating first module is specifically configured to calculate a euclidean distance between the weighted two-dimensional vector and the clustering center according to the following formula:
wherein, Fm，nFor any vector, V, in a weighted two-dimensional set of vectorspThe cluster center of the p-th class.
According to a specific implementation manner of the embodiment of the present invention, the fuzzy membership calculation module is configured to calculate a fuzzy membership of the kth weighted two-dimensional vector belonging to the pth cluster class:
wherein (F)m，n)kIs the k-th weighted two-dimensional vector; vpCluster centers for the p-th category; and c is the total number of the cluster categories.
According to a specific implementation manner of the embodiment of the present invention, the cluster center updating module is configured to calculate a cluster center of a pth class:
wherein f ism，nIs the two-dimensional vector; (u)m，n)kRepresenting the fuzzy membership degree of the kth weighted two-dimensional vector to each cluster class; a is a fuzzy weighting index; and N is the number of the two-dimensional vectors. p represents a class p category.
According to a specific implementation manner of the embodiment of the present invention, the objective function value calculating module is configured to calculate an objective function value:
wherein the content of the first and second substances,
(um，n)kfuzzy membership from the kth weighted two-dimensional vector to the pth clustering class; and N is the number of the two-dimensional vectors.
In a third aspect, an embodiment of the present invention provides an electronic device, where the electronic device includes: the device comprises a shell, a processor, a memory, a circuit board and a power circuit, wherein the circuit board is arranged in a space enclosed by the shell, and the processor and the memory are arranged on the circuit board; a power supply circuit for supplying power to each circuit or device of the electronic apparatus; the memory is used for storing executable program codes; the processor executes the program corresponding to the executable program code by reading the executable program code stored in the memory, and is used for executing the method of any one of the foregoing implementation modes.
In a fourth aspect, embodiments of the present invention also provide a computer-readable storage medium storing one or more programs, the one or more programs being executable by one or more processors to implement a method as described in any of the preceding implementations.
In this embodiment, a two-dimensional vector set generated by a self gray value and a neighborhood gray value of each pixel point in an image is weighted, a clustering center is initialized for the weighted two-dimensional vector, a fuzzy membership degree and a weighted euclidean distance are calculated, an objective function value can be further obtained, and if a difference between the objective function value and an objective function value after last iteration is smaller than an objective function change threshold, image segmentation is performed according to a maximum membership degree criterion. In the process of image segmentation, the relationship between the gray value of the image and the gray value of the neighborhood is considered, and the influence of the pixel information of the space on the image segmentation is expressed by adopting the weighted Euclidean distance, so that the image can be segmented more accurately, and a better segmentation effect is achieved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flowchart of an image segmentation method according to an embodiment of the present invention;
FIG. 2 is a grayscale histogram of an image;
FIG. 3 is a schematic structural diagram of an image segmentation apparatus according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an embodiment of an electronic device according to the present invention.
Detailed Description
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
It should be understood that the described embodiments are only some embodiments of the invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In a first aspect, an embodiment of the present invention provides an image segmentation method, which can segment an image more accurately.
Fig. 1 is a flowchart of an image segmentation method according to an embodiment of the present invention, as shown in fig. 1. The method of the embodiment may include:
In this embodiment, clustering groups a set of physical or abstract objects into classes composed of similar objects. In the image segmentation, the images can be classified into a class according to the similar gray values of all pixel points in the images, and the clustering class number can be classified into 2 classes, 3 classes, 5 classes and the like.
As an alternative embodiment, the cluster type may be determined based on a gray level histogram.
102, calculating the gray value of each pixel point in the image and the neighborhood gray value, and generating a two-dimensional vector set based on the gray value of each pixel point and the neighborhood gray value.
In this embodiment, the gray value represents the object by using black tone, that is, black is used as a reference color, and the image is displayed by black with different saturation degrees, and the range is generally from 0 to 255, white is 255, and black is 0; pixel, which refers to a minimum unit in an image represented by a sequence of numbers; the neighborhood gray value is a gray value corresponding to a pixel adjacent to any pixel point in the graph, for example, a gray value corresponding to one or more pixels in all pixel points adjacent to one pixel point in the graph. And generating a two-dimensional vector by the gray value corresponding to any pixel point in the image and the neighborhood gray value thereof, wherein a plurality of two-dimensional vectors can be generated by the gray values corresponding to all the pixel points in the image and the neighborhood gray values, all the two-dimensional vectors form a two-dimensional vector set, and the number of the two-dimensional vectors in the two-dimensional vector set is equal to the number of the gray values in the image.
And 103, weighting each two-dimensional vector in the two-dimensional vector set to obtain a weighted two-dimensional vector.
In the embodiment, the gray value of the image and the gray value of the neighborhood of the image both contribute to the image segmentation, and in the invention, the contribution of the gray value of the image to the image segmentation is considered to be 0, and the contribution of the neighborhood gray value to the image segmentation is used as a weighting coefficient to weight the two-dimensional vector. The weighting coefficients represent the contribution of the neighborhood gray value to the image segmentation. Because the influence of the neighborhood gray scale value of each pixel point in the image on the segmentation is different, the contribution of the neighborhood gray scale value corresponding to each pixel point to the image segmentation is different, namely the weighting coefficients are different.
The weighting coefficient comprehensively considers the influence of the spatial information around the pixel point on the image segmentation process, for example, for a certain pixel point, pixel points which greatly contribute to segmentation exist in the pixels of the four neighborhoods or the eight neighborhoods, and pixel points which greatly contribute to segmentation but have weakened clustering effect on the gray value of the pixel point exist.
And 104, initializing the clustering center based on the weighted two-dimensional vector.
In this embodiment, the cluster centers are randomly selected, and the number of the selected cluster centers is equal to the cluster category.
As an alternative embodiment, in the weighted two-dimensional vector set, the number of cluster centers is randomly selected to be equal to the cluster category.
And 105, calculating the Euclidean distance between the weighted two-dimensional vector and the clustering center.
In this embodiment, the weighted euclidean distance based on the weighting is calculated using the weighted two-dimensional vector.
And 106, calculating the fuzzy membership degree.
In this embodiment, the fuzzy membership degree refers to a degree that a certain weighted two-dimensional vector belongs to a certain cluster category.
And step 107, updating the clustering center according to the fuzzy membership.
In this embodiment, the clustering center of each clustering category is calculated according to the membership degree of the weighted two-dimensional vector to different clustering centers.
And 108, calculating an objective function value based on the fuzzy membership degree and the Euclidean distance between the weighted two-dimensional vector and a clustering center.
In this embodiment, after the fuzzy membership degrees from all weighted two-dimensional vectors to different clustering centers and the euclidean distances from all weighted two-dimensional vectors to different clustering centers are obtained through calculation, the objective function value is obtained through calculation.
In this embodiment, the objective function change threshold is preset to determine whether iterative computation of the fuzzy membership, the clustering center, and the euclidean distance between the weighted two-dimensional vector and the clustering center is required.
And step 110, if the difference value between the objective function value and the objective function value after the last iteration is smaller than the objective function change threshold, performing image segmentation according to the maximum membership criterion.
In the present embodiment, the maximum membership rule (maximum membership means principal) is the basic rule of fuzzy mathematicsOne of them is a direct method of model identification using fuzzy set theory, and for n actual models, it can be expressed as n fuzzy subsets A on the domain of discourse X1，A2，…，An，x0e.X is a specific identification object, if there is i0N or less, make Ai0(x0)＝max(A1(x0)，A2(x0)，...，An(x0) X) then call x0Relative membership to Ai0。
Performing image segmentation on each weighted two-dimensional vector according to the maximum membership criterion
And when the difference value between the objective function value and the objective function value after the last iteration is smaller than the objective function change threshold, determining which class the weighted two-dimensional vector belongs to according to the fuzzy membership and the clustering center of all weighted two-dimensional vectors obtained by the current iteration and the maximum value of the fuzzy membership from the weighted two-dimensional vectors to all classes, and further obtaining which class the gray value of the pixel point in the image belongs to.
In this embodiment, a two-dimensional vector set generated by a self gray value and a neighborhood gray value of each pixel point in an image is weighted, a clustering center is initialized for the weighted two-dimensional vector, a fuzzy membership degree and a weighted euclidean distance are calculated, an objective function value can be further obtained, and if a difference between the objective function value and an objective function value after last iteration is smaller than an objective function change threshold, image segmentation is performed according to a maximum membership degree criterion. In the process of image segmentation, the relationship between the gray value of the image and the adjacent gray value is considered, and the influence of the pixel information of the space on the image segmentation is represented by the weighted Euclidean distance, so that burrs, discrete points and noise points in the image segmentation result are fewer, the signal-to-noise ratio of the image is greatly improved, a better segmentation effect is obtained, and the image can be accurately segmented.
In an embodiment of the present invention, after determining whether a difference between the objective function value and the objective function value after the last iteration is smaller than an objective function change threshold (step 109), the method further includes:
and if the difference value between the objective function value and the objective function value after the last iteration is greater than or equal to the objective function change threshold, calculating the Euclidean distance between the weighted two-dimensional vector and the updated cluster center.
In this embodiment, if the difference between the objective function value and the objective function value after the last iteration is greater than or equal to the objective function change threshold, the euclidean distance between the weighted two-dimensional vector and the updated cluster center is recalculated using the cluster center obtained by the last iteration.
In this embodiment, if the difference between the objective function value and the objective function value after the last iteration is greater than or equal to the objective function change threshold, the cluster center obtained by the last iteration is used to recalculate the euclidean distance between the weighted two-dimensional vector and the updated cluster center, the fuzzy membership degree is updated again, the cluster center is updated again, so as to judge the objective function again, the iterative computations are performed for multiple times until the difference between the objective function value and the objective function value after the last iteration is less than the objective function change threshold, and the image segmentation is performed according to the maximum membership degree criterion, so that a better segmentation effect is obtained, and the image can be more accurately segmented.
In an embodiment of the present invention, after calculating the objective function value based on the fuzzy membership matrix and the weighted euclidean distance (step 108), the method further comprises:
and judging whether the iteration times reach a maximum iteration time threshold value, and if not, calculating the Euclidean distance between the weighted two-dimensional vector and the updated clustering center.
In this embodiment, the maximum iteration threshold is preset to determine whether iterative computation of the fuzzy membership, the clustering center, and the euclidean distance between the weighted two-dimensional vector and the clustering center is required.
In this embodiment, the image can be segmented quickly by setting the threshold of the maximum number of iterations.
In an embodiment of the present invention, calculating the neighborhood gray-scale value of each pixel point in the image includes:
the neighborhood gray value g (m, n) of each pixel point is calculated according to the following formula:
wherein s represents the width of a neighborhood window of a pixel point h (m, n);
m and n are coordinates of pixel points in the image, and i and j are relative coordinates of the pixel points in the width of the neighborhood window and the pixel points with the coordinates (m and n). h (m + i, n + j) is a pixel value with coordinates (m + i, n + j).
In this embodiment, the neighborhood gray value g (m, n) is a gray average value of all pixel points within the width of the neighborhood window, that is, an average neighborhood gray average value corresponding to a pixel point in the image is obtained, so that each two-dimensional vector in the two-dimensional vector set is generated by the gray value of the pixel point itself and the average neighborhood gray value.
In this embodiment, the average neighborhood gray value is obtained by averaging the gray values of all the pixels within the width of the neighborhood window, and the average neighborhood mean value is introduced to equalize or reduce the influence of the pixels with a larger difference from the gray values of the neighborhood pixels on the segmentation to a certain extent, so that the image can be segmented more accurately, and a good segmentation effect is obtained.
In an embodiment of the present invention, the calculating the euclidean distance between the weighted two-dimensional vector and the clustering center includes:
calculating the Euclidean distance between the weighted two-dimensional vector and the clustering center according to the following formula:
wherein, Fm，nFor any vector, V, in said set of weighted two-dimensional vectorspThe cluster center of the p-th class.
In this embodiment, the euclidean distance between the weighted two-dimensional vector and the clustering center is calculated, which is the euclidean distance based on the weighting.
In the embodiment, the image is divided by using the weighted Euclidean distance, so that the influence of noise on image division can be reduced, and a good division effect is obtained.
In an embodiment of the present invention, the calculating the fuzzy membership includes:
calculating the fuzzy membership degree of the kth weighted two-dimensional vector belonging to the pth clustering class according to the following formula:
wherein (F)m，n)kIs the k-th weighted two-dimensional vector; p is the pth cluster category; vpCluster centers for the p-th category; and c is the total number of the cluster categories.
In this embodiment, the fuzzy membership of the weighted two-dimensional vector belonging to each cluster category is obtained through calculation, and the fuzzy membership of all the weighted two-dimensional vectors belonging to each cluster category can be calculated according to a formula.
In this embodiment, the weighted two-dimensional vector is used to segment the image according to the fuzzy membership degree of each cluster category, and the deviation of various information on the segmentation is balanced, including the influence of the neighborhood pixel value, the burr, the discrete point and the noise point on the cluster category to which the pixel belongs, so that the image is segmented accurately, and a good segmentation effect is obtained.
In an embodiment of the present invention, the updating the clustering center according to the fuzzy membership includes:
calculating the cluster center of the p-type class category according to the following formula:
wherein f ism，nIs the two-dimensional vector; (u)m，n)kTo represent the k < th >The weighted two-dimensional vector reaches the fuzzy membership degree of each cluster category; a is a fuzzy weighting index; and N is the number of the two-dimensional vectors. p represents a class p category.
In this embodiment, the clustering center is recalculated according to the fuzzy membership, and the clustering center of each clustering category can be obtained according to the above formula.
In the embodiment, by recalculating the clustering centers, the segmentation of the image does not depend on the randomly generated clustering centers any more, and in the calculation process of the clustering centers, the influence of the neighborhood gray value on the image segmentation is considered, so that the clustering centers can be found more accurately.
In an embodiment of the present invention, the calculating an objective function value based on the fuzzy membership and the euclidean distance between the weighted two-dimensional vector and the clustering center includes:
the objective function value is calculated according to the following formula:
wherein the content of the first and second substances,
(um，n)kand the fuzzy membership degree from the kth weighted two-dimensional vector to the pth clustering class.
In this embodiment, on the basis of obtaining the euclidean distances and the fuzzy membership degrees between all the weighted two-dimensional vectors and the clustering centers through calculation, the objective function value can be obtained through calculation according to the formula of the objective function.
In this embodiment, the objective function value is obtained based on the euclidean distances and the fuzzy membership between all weighted two-dimensional vectors and the clustering center, and various information of deviation generated by image segmentation is comprehensively considered, so that the image can be segmented more accurately, and a better segmentation effect is obtained.
The following describes the technical solution of the embodiment of the method shown in fig. 1 in detail by using a specific embodiment.
Fig. 2 is a gray histogram of an image, and as shown in fig. 2, a specific image segmentation method is as follows:
step 1, determining the clustering category as 5 according to the gray level histogram.
And 2, calculating to obtain a certain gray value in the image as 60, and calculating the corresponding average neighborhood gray value as 80 by using the following formula.
Wherein s represents the width of a neighborhood window of the pixel point h (m, n), and s is taken to be 3; h (m, n) is 60, and h (m + i, n + j) is a gray scale value in the neighborhood of h (m, n).
Similarly, other gray values in the image and corresponding average neighborhood gray values are calculated, the gray values are used as horizontal coordinates, the average neighborhood gray values are used as vertical coordinates, two-dimensional vectors are generated, and the number of the generated two-dimensional vectors is 54000 because 54000 gray values exist in the original image.
Step 3, weighting the two-dimensional vectors (60, 80), wherein the contribution of the average neighborhood gray value to image segmentation is comprehensively considered by a weighting coefficient to be 2, and the weighted two-dimensional vectors are (120, 160); and weighting each two-dimensional vector by comprehensively considering the contribution of the average neighborhood gray value to image segmentation.
And 4, randomly selecting 5 two-dimensional vectors from the 54000 weighted two-dimensional vectors as cluster centers, wherein the first cluster center is (20, 30), and the other cluster centers are (55, 86), (110, 150), (170, 200), (220, 230) respectively.
And 5, calculating the distance between the two-dimensional vectors (120, 160) and the clustering centers (20, 30) by using the following formula:
step 6, calculating the fuzzy membership degree of the two-dimensional vectors (120, 160) belonging to the 1 st clustering class:
wherein p is 1, (F)m，n)k＝(120，160)，V1＝(20，30)、V2＝(55，86)、V3＝(110，150)、V4＝(170，200)、V5＝(220，230)，c＝5。
Similarly, fuzzy membership of the two-dimensional vector (120, 160) belonging to other classes and fuzzy membership of the other two-dimensional vector belonging to each class are calculated
Step 7, updating the clustering center according to the fuzzy membership degree calculated in the step 6,
calculating the cluster center of the p-type class category according to the following formula:
resulting in 5 updated cluster centers.
Step 8, calculating the objective function value according to the following formula:
wherein the content of the first and second substances,
(um，n)kand the fuzzy membership degree from the kth weighted two-dimensional vector to the pth clustering class.
Step 9, setting the maximum iteration threshold value as 100, calculating the Euclidean distance between the weighted two-dimensional vector and the updated clustering center under the condition that the maximum iteration threshold value does not reach 100 times, and repeating the step 6 to the step 9; when the number of calculations reaches 100 times, the next step is performed.
And step 10, segmenting the image according to the maximum membership criterion.
Fig. 3 is a schematic structural diagram of an image segmentation apparatus according to an embodiment of the present invention, and as shown in fig. 3, the apparatus according to the embodiment may include: a category number determining module 11, a vector set generating module 12, a weighting processing module 13, a cluster center initializing module 14, a Euclidean distance calculating first module 15, a fuzzy membership calculating module 16, a cluster center updating module 17, an objective function value calculating module 18, a first judging module 19 and an image segmentation module 20, wherein,
a category number determining module 11, configured to determine a clustering category number;
the vector set generation module 12 is configured to calculate a self gray value and a neighborhood gray value of each pixel point in the image, and generate a two-dimensional vector matrix based on the self gray value and the neighborhood gray value of each pixel point;
the weighting processing module 13 is configured to perform weighting processing on each two-dimensional vector in the two-dimensional vector set to obtain a weighted two-dimensional vector;
a cluster center initialization module 14, configured to perform initialization operation on a cluster center based on the weighted two-dimensional vector;
the Euclidean distance calculation first module 15 is used for calculating the Euclidean distance between the weighted two-dimensional vector and the clustering center;
a fuzzy membership degree calculating module 16, configured to calculate a fuzzy membership degree;
a cluster center updating module 17, configured to update the cluster center according to the fuzzy membership;
an objective function value calculation module 18, configured to calculate an objective function value based on the fuzzy membership and an euclidean distance between the weighted two-dimensional vector and a clustering center;
a first judging module 19, configured to judge whether a difference between the objective function value and the objective function value after the last iteration is smaller than an objective function change threshold;
and an image segmentation module 20, configured to perform image segmentation according to the maximum membership degree if a difference between the objective function value and the objective function value after the last iteration is smaller than a target function change threshold.
The apparatus of this embodiment may be used to implement the technical solution of the method embodiment shown in fig. 1, and the implementation principle and the technical effect are similar, which are not described herein again.
In an embodiment of the present invention, the apparatus further includes:
and a euclidean distance calculating second module 21, configured to calculate a euclidean distance between the weighted two-dimensional vector and the updated cluster center if a difference between the objective function value and the objective function value after the last iteration is greater than or equal to an objective function change threshold.
In an embodiment of the present invention, the apparatus further includes:
and the second judging module 22 is configured to judge whether the iteration number reaches a maximum iteration number threshold, and if not, calculate an euclidean distance between the weighted two-dimensional vector and the updated cluster center.
In an embodiment of the present invention, the vector matrix generating module includes:
a neighborhood gray calculating module 23, configured to calculate a neighborhood gray value g (m, n) of each pixel point according to the following formula:
wherein s represents the width of a neighborhood window of a pixel point h (m, n);
m and n are coordinates of pixel points in the image, and i and j are relative coordinates of the pixel points in the width of the neighborhood window and the pixel points with the coordinates (m and n). h (m + i, n + j) is a pixel value with coordinates (m + i, n + j).
In an embodiment of the present invention, the euclidean distance calculating module is specifically configured to calculate a euclidean distance between the weighted two-dimensional vector and the clustering center according to the following formula:
wherein, Fm，nFor any vector, V, in a weighted two-dimensional set of vectorspThe cluster center of the p-th class.
In an embodiment of the present invention, the fuzzy membership calculation module is configured to calculate a fuzzy membership of a kth weighted two-dimensional vector belonging to a pth cluster class:
wherein (F)m，n)kIs the k-th weighted two-dimensional vector; vpCluster centers for the p-th category; and c is the total number of the cluster categories.
In an embodiment of the present invention, the cluster center updating module is configured to calculate a cluster center of a pth class:
wherein f ism，nIs the two-dimensional vector; (u)m，n)kRepresenting the fuzzy membership degree of the kth weighted two-dimensional vector to each cluster class; a is a fuzzy weighting index; and N is the number of the two-dimensional vectors. p represents a class p category.
In an embodiment of the present invention, the objective function value calculating module is configured to calculate an objective function value:
wherein the content of the first and second substances,
(um，n)kfuzzy membership from the kth weighted two-dimensional vector to the pth clustering class; and N is the number of the two-dimensional vectors.
In a third aspect, an embodiment of the present invention further provides an electronic device, where the electronic device includes the apparatus in any of the foregoing embodiments.
Fig. 4 is a schematic structural diagram of an embodiment of an electronic device of the present invention, which can implement the process of the embodiment shown in fig. 1 of the present invention, and as shown in fig. 4, the electronic device may include: the device comprises a shell 41, a processor 42, a memory 43, a circuit board 44 and a power circuit 45, wherein the circuit board 44 is arranged inside a space enclosed by the shell 41, and the processor 42 and the memory 43 are arranged on the circuit board 44; a power supply circuit 45 for supplying power to each circuit or device of the electronic apparatus; the memory 43 is used for storing executable program code; the processor 42 executes a program corresponding to the executable program code by reading the executable program code stored in the memory 43, for executing the method described in any of the foregoing embodiments.
The specific execution process of the above steps by the processor 42 and the steps further executed by the processor 42 by running the executable program code may refer to the description of the embodiment shown in fig. 1 of the present invention, and are not described herein again.
The electronic device exists in a variety of forms, including but not limited to:
(1) ultra mobile personal computer device: the equipment belongs to the category of personal computers, has calculation and processing functions and generally has the characteristic of mobile internet access. Such terminals include: PDA, MID, and UMPC devices, etc., such as ipads.
(2) A server: the device for providing the computing service comprises a processor, a hard disk, a memory, a system bus and the like, and the server is similar to a general computer architecture, but has higher requirements on processing capacity, stability, reliability, safety, expandability, manageability and the like because of the need of providing high-reliability service.
(3) And other electronic equipment with data interaction function.
In a fourth aspect, embodiments of the present invention also provide a computer-readable storage medium storing one or more programs, the one or more programs being executable by one or more processors to implement a method as described in any of the preceding implementations.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. The term "comprising", without further limitation, means that the element so defined is not excluded from the group consisting of additional identical elements in the process, method, article, or apparatus that comprises the element.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments.
In particular, as for the apparatus embodiment, since it is substantially similar to the method embodiment, the description is relatively simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
For convenience of description, the above devices are described separately in terms of functional division into various units/modules. Of course, the functionality of the units/modules may be implemented in one or more software and/or hardware implementations of the invention.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (18)
1. A method of image segmentation, the method comprising:
determining the number of clustering categories;
calculating the self gray value and the neighborhood gray value of each pixel point in the image, and generating a two-dimensional vector set based on the self gray value and the neighborhood gray value of each pixel point;
weighting each two-dimensional vector in the two-dimensional vector set to obtain a weighted two-dimensional vector;
initializing a clustering center based on the weighted two-dimensional vector;
calculating the Euclidean distance between the weighted two-dimensional vector and the clustering center;
calculating fuzzy membership;
updating the clustering center according to the fuzzy membership degree;
calculating an objective function value based on the fuzzy membership degree and the Euclidean distance between the weighted two-dimensional vector and a clustering center;
judging whether the difference value between the objective function value and the objective function value after the last iteration is smaller than an objective function change threshold value or not;
and if the difference value between the objective function value and the objective function value after the last iteration is smaller than the objective function change threshold, carrying out image segmentation according to the maximum membership criterion.
2. The image segmentation method as claimed in claim 1, wherein after determining whether a difference between the objective function value and the objective function value after the last iteration is smaller than an objective function change threshold, the method further comprises:
and if the difference value between the objective function value and the objective function value after the last iteration is greater than or equal to the objective function change threshold, calculating the Euclidean distance between the weighted two-dimensional vector and the updated cluster center.
3. The image segmentation method according to claim 1, wherein after calculating an objective function value based on the fuzzy membership matrix and the weighted euclidean distance, the method further comprises:
and judging whether the iteration times reach a maximum iteration time threshold value, and if not, calculating the Euclidean distance between the weighted two-dimensional vector and the updated clustering center.
4. The image segmentation method according to claim 1, wherein calculating the neighborhood gray scale value of each pixel point in the image comprises:
the neighborhood gray value g (m, n) of each pixel point is calculated according to the following formula:
wherein s represents the width of a neighborhood window of a pixel point h (m, n);
m and n are coordinates of pixel points in the image, and i and j are relative coordinates of the pixel points in the width of the neighborhood window and the pixel points with the coordinates (m and n). h (m + i, n + j) is a pixel value with coordinates (m + i, n + j).
5. The image segmentation method according to claim 1, wherein the calculating the Euclidean distance between the weighted two-dimensional vector and the cluster center comprises:
calculating the Euclidean distance between the weighted two-dimensional vector and the clustering center according to the following formula:
wherein, Fm，nFor any vector, V, in said set of weighted two-dimensional vectorspThe cluster center of the p-th class.
6. The image segmentation method according to claim 1, wherein the calculating fuzzy membership comprises:
calculating the fuzzy membership degree of the kth weighted two-dimensional vector belonging to the pth clustering class according to the following formula:
wherein (F)m，n)kIs the k-th weighted two-dimensional vector; p is the pth cluster category; vpCluster centers for the p-th category; and c is the total number of the cluster categories.
7. The image segmentation method according to claim 1, wherein the updating the cluster center according to the fuzzy membership comprises:
calculating the cluster center of the p-type class category according to the following formula:
wherein f ism，nIs the two-dimensional vector; (u)m，n)kRepresenting the fuzzy membership degree of the kth weighted two-dimensional vector to each cluster class; a is a fuzzy weighting index; and N is the number of the two-dimensional vectors. p represents a class p category.
8. The image segmentation method according to claim 1, wherein the calculating an objective function value based on the fuzzy membership and a euclidean distance between the weighted two-dimensional vector and a cluster center comprises:
the objective function value is calculated according to the following formula:
wherein，(d′m，n)2＝(Fm，n-Vp)(Fm，n-Vp)T，
(um，n)kAnd the fuzzy membership degree from the kth weighted two-dimensional vector to the pth clustering class.
9. An image segmentation apparatus, comprising:
the category number determining module is used for determining the number of the clustering categories;
the vector set generation module is used for calculating the self gray value and the neighborhood gray value of each pixel point in the image and generating a two-dimensional vector matrix based on the self gray value and the neighborhood gray value of each pixel point;
the weighting processing module is used for weighting each two-dimensional vector in the two-dimensional vector set to obtain a weighted two-dimensional vector;
the cluster center initialization module is used for initializing the cluster center based on the weighted two-dimensional vector;
the Euclidean distance calculation first module is used for calculating the Euclidean distance between the weighted two-dimensional vector and the clustering center;
the fuzzy membership calculation module is used for calculating fuzzy membership;
the cluster center updating module is used for updating the cluster center according to the fuzzy membership degree;
the objective function value calculation module is used for calculating an objective function value based on the fuzzy membership degree and the Euclidean distance between the weighted two-dimensional vector and a clustering center;
the first judgment module is used for judging whether the difference value between the objective function value and the objective function value after the last iteration is smaller than a target function change threshold value or not;
and the image segmentation module is used for segmenting the image according to the maximum membership degree if the difference value between the objective function value and the objective function value after the last iteration is smaller than the objective function change threshold value.
10. The image segmentation apparatus according to claim 9, wherein the apparatus further comprises:
and the Euclidean distance calculation second module is used for calculating the Euclidean distance between the weighted two-dimensional vector and the updated clustering center if the difference value between the objective function value and the objective function value after the last iteration is greater than or equal to the objective function change threshold value.
11. The image segmentation apparatus according to claim 9, wherein the apparatus further comprises:
and the second judgment module is used for judging whether the iteration times reach a maximum iteration time threshold value, and if not, calculating the Euclidean distance between the weighted two-dimensional vector and the updated clustering center.
12. The image segmentation apparatus as set forth in claim 9, wherein the vector matrix generation module includes:
the neighborhood gray level calculating module is used for calculating the neighborhood gray level value g (m, n) of each pixel point according to the following formula:
wherein s represents the width of a neighborhood window of a pixel point h (m, n);
m and n are coordinates of pixel points in the image, and i and j are relative coordinates of the pixel points in the width of the neighborhood window and the pixel points with the coordinates (m and n). h (m + i, n + j) is a pixel value with coordinates (m + i, n + j).
13. The image segmentation device according to claim 9, wherein the euclidean distance calculation first module is specifically configured to calculate the euclidean distance between the weighted two-dimensional vector and the cluster center according to the following formula:
wherein, Fm，nFor any vector, V, in a weighted two-dimensional set of vectorspThe cluster center of the p-th class.
14. The image segmentation apparatus as claimed in claim 9, wherein the fuzzy membership calculation module is configured to calculate a fuzzy membership of the kth weighted two-dimensional vector belonging to the pth cluster class:
wherein (F)m，n)kIs the k-th weighted two-dimensional vector; vpCluster centers for the p-th category; and c is the total number of the cluster categories.
15. The image segmentation device according to claim 9, wherein the cluster center update module is configured to calculate a cluster center of a p-type class:
wherein f ism，nIs the two-dimensional vector; (u)m，n)kRepresenting the fuzzy membership degree of the kth weighted two-dimensional vector to each cluster class; a is a fuzzy weighting index; and N is the number of the two-dimensional vectors. p represents a class p category.
16. The image segmentation device as claimed in claim 9, wherein the objective function value calculation module is configured to calculate an objective function value:
wherein, (d'm，n)2＝(Fm，n-Vp)(Fm，n-Vp)T，
(um，n)kFuzzy membership from the kth weighted two-dimensional vector to the pth clustering class; and N is the number of the two-dimensional vectors.
17. An electronic device, characterized in that the electronic device comprises: the device comprises a shell, a processor, a memory, a circuit board and a power circuit, wherein the circuit board is arranged in a space enclosed by the shell, and the processor and the memory are arranged on the circuit board; a power supply circuit for supplying power to each circuit or device of the electronic apparatus; the memory is used for storing executable program codes; the processor executes a program corresponding to the executable program code by reading the executable program code stored in the memory for performing the method of any of the preceding claims.
18. A computer readable storage medium, characterized in that the computer readable storage medium stores one or more programs which are executable by one or more processors to implement the method of any preceding claim.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811632456.4A CN110866527A (en) | 2018-12-28 | 2018-12-28 | Image segmentation method and device, electronic equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811632456.4A CN110866527A (en) | 2018-12-28 | 2018-12-28 | Image segmentation method and device, electronic equipment and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110866527A true CN110866527A (en) | 2020-03-06 |
Family
ID=69651898
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811632456.4A Pending CN110866527A (en) | 2018-12-28 | 2018-12-28 | Image segmentation method and device, electronic equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110866527A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112084988A (en) * | 2020-06-08 | 2020-12-15 | 深圳佑驾创新科技有限公司 | Lane line instance clustering method and device, electronic equipment and storage medium |
| CN114520894A (en) * | 2020-11-18 | 2022-05-20 | 成都极米科技股份有限公司 | Projection area determining method and device, projection equipment and readable storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101894372A (en) * | 2010-08-03 | 2010-11-24 | 新疆大学 | New noise-containing remote sensing image segmentation method |
| CN103366367A (en) * | 2013-06-19 | 2013-10-23 | 西安电子科技大学 | Pixel number clustering-based fuzzy C-average value gray level image splitting method |
| CN103839269A (en) * | 2014-03-21 | 2014-06-04 | 南京大学 | Image segmentation method based on quaternion and fuzzy C-means clustering |
| CN104156943A (en) * | 2014-07-14 | 2014-11-19 | 西安电子科技大学 | Multi-target fuzzy cluster image variance detecting method based on non-control-neighborhood immune algorithm |
| CN105976373A (en) * | 2016-05-05 | 2016-09-28 | 江南大学 | Kernel fuzzy C-means image segmentation algorithm based on neighborhood information entropy |
| CN106127735A (en) * | 2016-06-14 | 2016-11-16 | 中国农业大学 | A kind of facilities vegetable edge clear class blade face scab dividing method and device |
| CN107169962A (en) * | 2017-05-16 | 2017-09-15 | 西安电子科技大学 | The gray level image fast partition method of Kernel fuzzy clustering is constrained based on space density |
| CN108830289A (en) * | 2018-04-28 | 2018-11-16 | 河南师范大学 | A kind of image clustering method and device based on improved fuzzy C-means clustering |
-
2018
- 2018-12-28 CN CN201811632456.4A patent/CN110866527A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101894372A (en) * | 2010-08-03 | 2010-11-24 | 新疆大学 | New noise-containing remote sensing image segmentation method |
| CN103366367A (en) * | 2013-06-19 | 2013-10-23 | 西安电子科技大学 | Pixel number clustering-based fuzzy C-average value gray level image splitting method |
| CN103839269A (en) * | 2014-03-21 | 2014-06-04 | 南京大学 | Image segmentation method based on quaternion and fuzzy C-means clustering |
| CN104156943A (en) * | 2014-07-14 | 2014-11-19 | 西安电子科技大学 | Multi-target fuzzy cluster image variance detecting method based on non-control-neighborhood immune algorithm |
| CN105976373A (en) * | 2016-05-05 | 2016-09-28 | 江南大学 | Kernel fuzzy C-means image segmentation algorithm based on neighborhood information entropy |
| CN106127735A (en) * | 2016-06-14 | 2016-11-16 | 中国农业大学 | A kind of facilities vegetable edge clear class blade face scab dividing method and device |
| CN107169962A (en) * | 2017-05-16 | 2017-09-15 | 西安电子科技大学 | The gray level image fast partition method of Kernel fuzzy clustering is constrained based on space density |
| CN108830289A (en) * | 2018-04-28 | 2018-11-16 | 河南师范大学 | A kind of image clustering method and device based on improved fuzzy C-means clustering |
Non-Patent Citations (3)
| Title |
|---|
| 刘键庄: "基于二维直方图的图象模糊聚类分割方法", 《电子学报》 * |
| 李伟: "基于模糊C均值聚类的图像分割算法研究", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
| 熊和金 等: "《智能信息处理 第2版》", 31 August 2012 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112084988A (en) * | 2020-06-08 | 2020-12-15 | 深圳佑驾创新科技有限公司 | Lane line instance clustering method and device, electronic equipment and storage medium |
| CN112084988B (en) * | 2020-06-08 | 2024-01-05 | 武汉佑驾创新科技有限公司 | Lane line instance clustering method and device, electronic equipment and storage medium |
| CN114520894A (en) * | 2020-11-18 | 2022-05-20 | 成都极米科技股份有限公司 | Projection area determining method and device, projection equipment and readable storage medium |
| CN114520894B (en) * | 2020-11-18 | 2022-11-15 | 成都极米科技股份有限公司 | Projection area determining method and device, projection equipment and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108171663B (en) | Image filling system of convolutional neural network based on feature map nearest neighbor replacement | |
| CN110648289B (en) | Image noise adding processing method and device | |
| CN112149705A (en) | Method and system for training classification model, computer equipment and storage medium | |
| US9619729B2 (en) | Density measuring device, density measuring method, and computer program product | |
| Wang et al. | Energy based competitive learning | |
| CN109685830B (en) | Target tracking method, device and equipment and computer storage medium | |
| CN109658378B (en) | Pore identification method and system based on soil CT image | |
| CN111223128A (en) | Target tracking method, device, equipment and storage medium | |
| CN112149825A (en) | Neural network model training method and device, electronic equipment and storage medium | |
| JP2015036939A (en) | Feature extraction program and information processing apparatus | |
| CN110866527A (en) | Image segmentation method and device, electronic equipment and readable storage medium | |
| CN114091594A (en) | Model training method and device, equipment and storage medium | |
| CN114118259A (en) | Target detection method and device | |
| CN116109907B (en) | Target detection method, target detection device, electronic equipment and storage medium | |
| CN113011532A (en) | Classification model training method and device, computing equipment and storage medium | |
| CN112818946A (en) | Training of age identification model, age identification method and device and electronic equipment | |
| Tang et al. | Boundary-guided optimization framework for saliency refinement | |
| CN116977271A (en) | Defect detection method, model training method, device and electronic equipment | |
| Liu et al. | Ranking-preserving cross-source learning for image retargeting quality assessment | |
| CN110717037A (en) | Method and device for classifying users | |
| US20220092448A1 (en) | Method and system for providing annotation information for target data through hint-based machine learning model | |
| CN114255381A (en) | Training method of image recognition model, image recognition method, device and medium | |
| JP4879257B2 (en) | Moving object tracking device, moving object tracking method, and moving object tracking program | |
| CN114359936A (en) | Answer sheet filling identification method, model construction method, equipment and storage medium | |
| CN111815658B (en) | Image recognition method and device |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200306 |