CN115131588B - Image robust clustering method based on fuzzy clustering - Google Patents
Image robust clustering method based on fuzzy clustering Download PDFInfo
- Publication number
- CN115131588B CN115131588B CN202210665911.0A CN202210665911A CN115131588B CN 115131588 B CN115131588 B CN 115131588B CN 202210665911 A CN202210665911 A CN 202210665911A CN 115131588 B CN115131588 B CN 115131588B
- Authority
- CN
- China
- Prior art keywords
- clustering
- image
- sample
- data
- noise
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 239000011159 matrix material Substances 0.000 claims abstract description 39
- 238000005457 optimization Methods 0.000 claims description 20
- 230000001629 suppression Effects 0.000 claims description 13
- 238000012216 screening Methods 0.000 claims description 6
- 238000012163 sequencing technique Methods 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 abstract description 25
- 238000004364 calculation method Methods 0.000 abstract description 5
- 230000006870 function Effects 0.000 description 18
- 230000000694 effects Effects 0.000 description 4
- 238000010606 normalization Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001351 cycling effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
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
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
- G06V10/763—Non-hierarchical techniques, e.g. based on statistics of modelling distributions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Multimedia (AREA)
- Computational Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Data Mining & Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Algebra (AREA)
- Probability & Statistics with Applications (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to an image robust clustering method based on fuzzy clustering, which screens out images polluted by noise while clustering image data, reserves pure pollution-free image data and has higher robustness to noise change. The invention is an unsupervised algorithm, does not need to use tag data, and reduces the time for acquiring a large amount of tag data. The algorithm does not need to update the graph matrix in the solving process, so that the calculation complexity of the algorithm is reduced, and the calculation speed is increased. Thus, a fast, efficient and robust clustering of noisy contaminated images can be achieved. The regularization parameter of the objective function can be optimized through iteration, the corresponding regularization parameter of each sample can be calculated in a self-adaptive mode, difficulty in adjusting the regularization parameter is greatly reduced in the application process, labor cost is saved, noise-polluted image data is screened out in a robust mode, and image clustering accuracy is improved.
Description
Technical Field
The invention belongs to the fields of image recognition, classification and pattern recognition, and relates to an image robust clustering method based on fuzzy clustering.
Background
With the development of computer technology and digital imaging systems, it is becoming more and more convenient for people to transfer information through images. However, in a real environment, image information is easily polluted by noise, so that the image quality is lost to a certain extent, and effective identification of images is difficult. Because the processed image information is more and more complex and the tag acquisition difficulty is more and more high, the unsupervised image clustering technology is widely focused on the application of the unsupervised image clustering technology in the information age, and the image clustering technology can cluster the images in the image database according to the similarity, so that the similarity of the images in the same cluster is as large as possible and the similarity between different clusters is as small as possible. The image information is easy to be affected by noise, and if the image polluted by the noise is still subjected to traditional unsupervised image clustering, the accuracy and the reliability of image retrieval can be greatly affected. And screening out images polluted by noise, and comparing the new images with clusters with higher similarity in a database one by one to quickly finish identification and classification. Therefore, noise-suppressed clustering of image data prior to image retrieval can effectively and quickly achieve high-quality image data retrieval.
Li Kang et al (Chinese scientific and technological paper of a hyperspectral image classification-oriented fuzzy spectral clustering algorithm, 2021,16 (07): 743-747.) a Fuzzy Similarity Metric Spectral Clustering (FSMSC) algorithm for hyperspectral remote sensing image classification aims at constructing an effective fuzzy similarity matrix by introducing a fuzzy similarity metric and a stable anchor graph structure and improving the performance of the clustering algorithm. However, updating the graph matrix also increases the time complexity of the fuzzy clustering algorithm, affecting the operation speed. Although graph learning and fuzzy clustering learning are integrated into a joint learning framework, the method is limited by a traditional fuzzy clustering algorithm, has poor robustness, cannot effectively remove noise data, and influences subsequent image data retrieval.
Disclosure of Invention
Technical problem to be solved
In order to avoid the defects of the prior art, the invention provides an image robust clustering method based on fuzzy clustering, which aims at solving the problems that the existing supervised image algorithm needs to consume a large amount of time to acquire data labels, and noise cannot be effectively solved to influence image data retrieval, and the robustness is poor.
Technical proposal
The image robust clustering method based on fuzzy clustering is characterized by comprising the following steps:
step 1: for n pieces of picture data with the resolution of u×v, elongating each picture to obtain a row vector of 1×d, where d=u×v; converting image data composed of n pictures into target data matrixWherein each row of the matrix is>Representing an image; each sheet ofThe image represents a data sample; giving the true class number c contained in the target data, randomly initializing the clustering centers of c clusters to obtain initial +.>Is the centroid of the j-th cluster;
step 2: establishing robust fuzzy clustering RFCM framework for noise suppression
Wherein the method comprises the steps ofIs fuzzy membership, and the centroid matrix of the cluster is +.>The matrix Y represents each element Y ij Representing the membership degree of the ith sample belonging to the jth cluster; />For screening n-k noise data s i Values for the ith element of s;
step 3: alternating iterative optimization robust fuzzy clustering RFCM framework
The solving steps are as follows:
step 3.1: randomly initializing the cluster centers of c clusters to obtain an initial cluster centerInitializing all elements in Y to Y ij =1/c; definition e i To a certain membership degree to sample x i Weighting and summing the obtained values to all cluster center distances, and adding e i Arranged in order from small to large as e 1 ≤e 2 ≤...≤e k ≤...≤e n Calculating to obtain the corresponding s i ；
Step 3.2: different optimizations are respectively carried out on the real sample and the noise data to obtain Y
The first k samples nearest to all cluster centers correspond to s i =1, s corresponding to the remaining samples i =0; e after the sequence of the two steps i Corresponding sample x i Sequencing to obtain a sequenced data matrixCorresponding to the membership matrix after sequencing>
Step 3.2.1: when the sample corresponds to s i When=1, the q clusters nearest to the real sample, the membership matrix is thinned, and y is limited i L of (2) 0 The norm is q, defined asThe i-th real sample point->The distance squared to the j-th cluster center, the RFCM frame equivalent translates to
The parameter gamma in the objective function is used for adaptively calculating the optimal parameter gamma;
optimizing to obtain the optimal membership degree corresponding to the selected real samplei∈{1,2,...,k}
Step 3.2.2: for noise data in the optimization process, when the sample corresponds to s i When=0, the optimal membership value corresponding to the noise data in the optimization process is obtained by solving the cauchy inequality
Obtaining the optimal solution of all sample point membership degreesIs that
Step 3.3: fixing s and Y, obtaining the sub-problem of the RFCM framework
The sub-problem is solved by solving the m bias guide equal to 0:
in each optimization process, m, Y and s are continuously updated, and the next iteration operation is carried out again until m is not changed; one row of sample data corresponds to one picture, and according to the obtained membership matrix Y, the label of the cluster corresponding to the maximum value of each row is selected as the classified type of the picture, namely a predicted label vector is obtained, and image clustering cluster division is realized; the obtained s vector contains k 1, n-k 0 s corresponding to the ith image i =0, then this pictureIs screened as an image contaminated with noise.
Advantageous effects
According to the image robust clustering method based on fuzzy clustering, images polluted by noise are screened out while image data are clustered, pure pollution-free image data are reserved, and the robustness to noise change is high. The invention is an unsupervised algorithm, does not need to use tag data, and reduces the time for acquiring a large amount of tag data. The algorithm does not need to update the graph matrix in the solving process, so that the calculation complexity of the algorithm is reduced, and the calculation speed is increased. Thus, a fast, efficient and robust clustering of noisy contaminated images can be achieved.
The invention provides an improved noise suppression image clustering method of a robust fuzzy clustering framework based on FCM, which can be used for both robustly screening noise polluted image data and clustering images in the application process, and simultaneously improves the robustness and precision of an algorithm and the data processing speed of sparsification processing of a membership matrix. In the algorithm, the objective function consists of a robust noise suppression term and a regularization term, and an adaptive weight is added to each sample point through iterative optimization of the objective function, so that a clean sample and a noise pollution sample are screened, and the robustness of the algorithm is enhanced. The noise pollution data samples can be screened by optimizing the objective function, and then the image clustering of noise suppression is carried out.
The method provided by the invention has the beneficial effects that:
(1) A robust fuzzy clustering algorithm is provided, in the algorithm, a robust noise suppression term of an objective function enables the algorithm to add self-adaptive weights to each sample point through iterative optimization of the objective function, and therefore pure samples and noise pollution samples are screened, and robustness of the algorithm is enhanced.
(2) According to the noise suppression image clustering method based on the robust fuzzy clustering, the membership matrix is thinned while the robust noise suppression is carried out, so that a more effective sample and feature distribution structure is obtained, the influence of noise pollution on image data clustering is avoided, the storage amount of data is reduced, the calculated amount of the data is reduced, and the calculation efficiency is improved.
(3) The regularization parameter of the objective function can be optimized through iteration, the corresponding regularization parameter of each sample can be calculated in a self-adaptive mode, difficulty in adjusting the regularization parameter is greatly reduced in the application process, labor cost is saved, noise-polluted image data is screened out in a robust mode, and image clustering accuracy is improved.
Drawings
Fig. 1: is a method flow chart
Fig. 2: is an example of a partially noisy contaminated image
Fig. 3: is a graph of the detection results of the method on a specific data set
Detailed Description
The invention will now be further described with reference to examples, figures:
the invention is realized by the following technical scheme, and the method for clustering the noise-suppressed image based on the robust fuzzy clustering comprises the following specific steps:
step 1: obtaining image data information to construct a data matrix
For n pieces of picture data with the resolution of u×v, elongating each picture to obtain a row vector of 1×d, where d=u×v; converting image data composed of n pictures into target data matrixWherein each row of the matrix is>Representing an image; each image represents a data sample; giving the true class number c contained in the target data, randomly initializing the clustering centers of c clusters to obtain initial +.>Is the centroid of the j-th cluster.
Step 2: establishing a Robust Fuzzy Clustering (RFCM) framework that can perform noise suppression
The Fuzzy C-means (FCM) algorithm is short for FCM algorithm, and the frame introduces noise suppression term and regularization term for improving noise suppression robustness.
Wherein the method comprises the steps ofIs fuzzy membership, and the centroid matrix of the cluster is +.>The matrix Y represents each element Y ij Indicating the membership of the ith sample to the jth cluster. />For screening n-k noise data s i Is the value of the ith element of s.
Step 3: alternately iterating and optimizing an objective function:
solving three variables of m, Y and s in an objective function by adopting an alternate iterative optimization method, initializing m and Y at first, and calculating to obtain s according to a formula; fixing s and m, and respectively carrying out different optimization on the real sample and noise data to obtain Y; then fixing s and Y, solving m according to a formula, and sequentially cycling until convergence;
the solving steps are as follows:
step 3.1: according to the clustering center of the randomly initialized c clusters, obtaining an initial clusterInitializing all elements in Y to Y ij =1/c. Consider an initial probability of equalizing the sample distribution across each class. Definition e i To a certain membership degree to sample x i The distances to the centers of all clusters are weighted and the resulting values summed. Will e i Arranged in order from small to large as e 1 ≤e 2 ≤...≤e k ≤…≤e n The corresponding s can be calculated i
Step 3.2: and carrying out different optimization on the real sample and the noise data to obtain Y.
In step 3.1 we get the s corresponding to the first k samples nearest to all cluster centers i =1, s corresponding to the remaining samples i =0. Will e i Sorting the samples x corresponding to the samples x from small to large i Sorting is carried out to obtain a sorted data matrixCorresponding to the membership matrix after sequencing>
Step 3.2.1: for real samples. When the sample corresponds to s i When=1, only consider q clusters nearest to the real sample, sparsify the membership matrix, limit y i L of (2) 0 The norm is q. Definition of the definitionFor the i-th real sample point selected +.>The problem can be equivalently translated into the square of the center distance to the j cluster
The parameter gamma in the objective function generally needs to be properly adjusted to avoid the occurrence of trivial solutions, and the present invention can adaptively calculate the optimal parameter gamma.
Optimizing to obtain the optimal membership degree corresponding to the selected real samplei∈{1,2,...,k}
Step 3.2.2: for noise data in the optimization process. When the sample corresponds to s i When=0, the optimal membership value corresponding to the noise data in the optimization process is obtained by solving the cauchy inequality
The optimal solution of all sample point membership degree can be obtainedIs that
Step 3.3:
the partial derivative of the m of the sub-problem pair of the objective function is equal to 0, and can be solved
And finally, after m, Y and s are updated, the next iteration operation is carried out again until m is not changed. And selecting a label of a cluster corresponding to the maximum value of each row of the obtained membership matrix Y as the classified type of the picture, namely obtaining a predicted label vector, and realizing image clustering cluster division. Obtained byThe s vector of (2) contains k 1, n-k 0 s, and the s corresponding to the ith image i =0, then this picture is screened as an image contaminated with noise.
Specific examples:
the comprehensive model solving process of the noise suppression image clustering method based on the robust fuzzy clustering is shown in figure 1, an ORL face image dataset is selected for clustering examples, the ORL face image dataset contains 400 face images in total, the resolution is 92 multiplied by 112, each face image corresponds to one sample, and the real labels corresponding to the 400 images areThe predictive label of 400 images obtained by the clustering algorithm is +.>Wherein the real label z t Only for final cluster effect verification, and is not included in the clustering algorithm itself. To test the algorithm, taking the example of adding 40% proportion of noise, the number of noise contaminated image samples p=160. The specific embodiment comprises the following steps:
step one, input ORL data matrixA true class number c, a noise contaminated image sample number p=160 and a membership degree sparsification parameter q, wherein each row of the matrix +.>For one sample, n=400 is the number of samples, d=92×112=10304 is the dimension of the data matrix, and c=40 is the number of real categories contained in the face data.
And randomly initializing the cluster centers of the c clusters to obtain the initial m. Initializing all elements in Y to Y ij =1/c=1/40, k=n-p=240. Consider an initial probability of equalizing the sample distribution across each class. M and Y can be fixed, and s can be calculated. Fixing m and Y, the objective function is converted into
Definition e i To a certain membership degree to sample x i The distances to the centers of all clusters are weighted and the resulting values summed.
Will e i Arranged in order from small to large as e 1 ≤e 2 ≤...≤e k ≤…≤e n The optimal solution of the objective function in the constraint s can be calculated T Its corresponding s under the condition 1=k i
According to the method, sample data which is not polluted by noise and samples polluted by noise can be obtained through screening, and the sample data and the samples polluted by noise are continuously optimized in the subsequent iteration process.
Step two: and carrying out different optimization on the real sample and the noise data to obtain Y.
In step one, s corresponding to the first k samples nearest to all cluster centers are obtained i =1, s corresponding to the remaining samples i =0. Will e i Sorting the samples x corresponding to the samples x from small to large i Sorting is carried out to obtain a sorted data matrixCorresponding to the membership matrix after sequencing>
Step 2.1: for real samples. Definition of the definitionIs the ordered data matrix +.>A vector of elements of row i. />Is a membership matrix after sequencing>The j-th element of the i-th row. When the sample corresponds to s i When=1, the sub-problem of the objective function is that
Considering only q clusters nearest to the real sample, sparsifying the membership matrix, and limiting y i L of (2) 0 The norm is q. Definition of the definitionFor the i-th real sample point selected +.>The problem can be equivalently translated into the square of the center distance to the j cluster
The second term of the original objective function is a regularization term to avoid the trivial solution of two extremes: only nearest neighbor samples have a similarity of 1 and all samples have a similarity of 1/n. By Lagrangian multiplier method, KKT conditions and constraints
Since each term of the problem is independent for i, we can be for eachSeparately solving, namely solving an objective function equivalent sub-problem by using a Lagrange multiplier method, and optimizing to obtain the optimal membership degree corresponding to the selected real sample>i∈{1,2,...,k}
Step 2.2: for noise data in the optimization process. When the sample corresponds to s i When=0, the optimal membership value corresponding to the noise data in the optimization process is obtained by solving the cauchy inequality
The optimal solution of all sample point membership degree can be obtainedIs that
Step three: fixing s and Y, solving m
At this time, the objective function subproblem is rewritten as
The partial derivative of the m of the sub-problem pair of the objective function is equal to 0, and can be solved
So far, s, Y and m are updated completely, and next iteration operation is carried out again until the cluster center m is not updated any more, namely the change of the cluster center m is smaller than a certain threshold value. Contaminated sample corresponds to s i After the completion of the solution, the optimized s is obtained i The ith sample corresponding to=0 is screened as noise-contaminated image samples, and 160 noise-contaminated image samples can be screened in total. Finally, clustering prediction labels capable of directly obtaining 400 face imagesSo as to achieve a considerable clustering effect. Different from classification, the predictive labels obtained by the clustering method can only achieve grouping effect under the condition of no supervision, so that the numbers in the predictive labels are in one-to-one correspondence with the real class labels, but the specific correspondence cannot be known, and therefore, the face data images subjected to no supervision grouping can be used for grouping and classifying the face data images without label information, so as to assist the face retrieval and greatly improve the retrieval precision and speed. By comparing true tags z t And a predictive label z obtained by a clustering algorithm p The accuracy of the image clusters is calculated.
Taking the ORL face image dataset (400 pictures, each with 92×112 pixels) as an example. When 40% noise is added, the clustering accuracy of the FCM on ORL face image data is only 8.61%, and the clustering normalization mutual information is 14.13%. When 40% of noise is added, the clustering accuracy of the noise suppression image clustering (RFCM) based on the robust fuzzy clustering on the ORL face image data set is 64.83%, clustering normalization mutual information is 71.29%, 56.22% and 57.16% of clustering accuracy of the face image data are respectively improved, and the clustering accuracy of the face image data is remarkably improved.
Claims (1)
1. The image robust clustering method based on fuzzy clustering is characterized by comprising the following steps:
step 1: for n-sheet uXv resolutionDrawing each picture to obtain a row vector of 1×d, wherein d=u×v; converting image data composed of n pictures into target data matrixWherein each row of the matrix is>Representing an image; each image represents a data sample; giving the true class number c contained in the target data, randomly initializing the clustering centers of c clusters to obtain initial +.>Is the centroid of the j-th cluster;
step 2: establishing robust fuzzy clustering RFCM framework for noise suppression
Wherein the method comprises the steps ofIs fuzzy membership, and the centroid matrix of the cluster is +.>The matrix Y represents each element Y ij Representing the membership degree of the ith sample belonging to the jth cluster; />For screening n-k noise data s i Values for the ith element of s;
step 3: alternating iterative optimization robust fuzzy clustering RFCM framework
The solving steps are as follows:
step 3.1: randomly initializing the cluster centers of c clusters to obtain an initial clusterA kind of electronic deviceInitializing all elements in Y to Y ij =1/c; definition e i To a certain membership degree to sample x i Weighting and summing the obtained values to all cluster center distances, and adding e i Arranged in order from small to large as e 1 ≤e 2 ≤...≤e k ≤...≤e n Calculating to obtain the corresponding s i ；
Step 3.2: different optimizations are respectively carried out on the real sample and the noise data to obtain Y
The first k samples nearest to all cluster centers correspond to s i =1, s corresponding to the remaining samples i =0; e after the sequence of the two steps i Corresponding sample x i Sequencing to obtain a sequenced data matrixCorresponding to the ordered membership matrix
Step 3.2.1: when the sample corresponds to s i When=1, the q clusters nearest to the real sample, the membership matrix is thinned, and y is limited i L of (2) 0 The norm is q, defined asThe i-th real sample point->The distance squared to the j-th cluster center, the RFCM frame equivalent translates to
The parameter gamma in the objective function is used for adaptively calculating the optimal parameter gamma;
optimizing to obtain the optimal membership degree corresponding to the selected real sample
Step 3.2.2: for noise data in the optimization process, when the sample corresponds to s i When=0, the optimal membership value corresponding to the noise data in the optimization process is obtained by solving the cauchy inequality
Obtaining the optimal solution of all sample point membership degreesIs that
Step 3.3: fixing s and Y, obtaining the sub-problem of the RFCM framework
The sub-problem is solved by solving the m bias guide equal to 0:
in each optimization process, m, Y and s are continuously updated, and the next iteration operation is carried out again until m is not changed; one row of sample data corresponds to one picture, and according to the obtained membership matrix Y, the label of the cluster corresponding to the maximum value of each row is selected as the classified type of the picture, namely a predicted label vector is obtained, and image clustering cluster division is realized; the obtained s vector contains k 1, n-k 0 s corresponding to the ith image i =0, then this picture is screened as an image contaminated with noise.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210665911.0A CN115131588B (en) | 2022-06-13 | 2022-06-13 | Image robust clustering method based on fuzzy clustering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210665911.0A CN115131588B (en) | 2022-06-13 | 2022-06-13 | Image robust clustering method based on fuzzy clustering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115131588A CN115131588A (en) | 2022-09-30 |
| CN115131588B true CN115131588B (en) | 2024-02-23 |
Family
ID=83377270
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210665911.0A Active CN115131588B (en) | 2022-06-13 | 2022-06-13 | Image robust clustering method based on fuzzy clustering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115131588B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101866522B1 (en) * | 2016-12-16 | 2018-06-12 | 인천대학교 산학협력단 | Object clustering method for image segmentation |
| CN110211126A (en) * | 2019-06-12 | 2019-09-06 | 西安邮电大学 | Image partition method based on intuitionistic fuzzy C mean cluster |
| WO2021007744A1 (en) * | 2019-07-15 | 2021-01-21 | 广东工业大学 | Kernel fuzzy c-means fast clustering algorithm with integrated spatial constraints |
| CN113538445A (en) * | 2021-07-30 | 2021-10-22 | 安徽工业大学 | Image segmentation method and system based on weighted robust FCM clustering |
-
2022
- 2022-06-13 CN CN202210665911.0A patent/CN115131588B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101866522B1 (en) * | 2016-12-16 | 2018-06-12 | 인천대학교 산학협력단 | Object clustering method for image segmentation |
| CN110211126A (en) * | 2019-06-12 | 2019-09-06 | 西安邮电大学 | Image partition method based on intuitionistic fuzzy C mean cluster |
| WO2021007744A1 (en) * | 2019-07-15 | 2021-01-21 | 广东工业大学 | Kernel fuzzy c-means fast clustering algorithm with integrated spatial constraints |
| CN113538445A (en) * | 2021-07-30 | 2021-10-22 | 安徽工业大学 | Image segmentation method and system based on weighted robust FCM clustering |
Non-Patent Citations (1)
| Title |
|---|
| 吴成茂 ; 吴其平 ; .一种基于改进PFCM的鲁棒图像分割算法.西安邮电大学学报.2017,(05),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115131588A (en) | 2022-09-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111860612B (en) | Unsupervised hyperspectral image hidden low-rank projection learning feature extraction method | |
| CN109961089B (en) | Small sample and zero sample image classification method based on metric learning and meta learning | |
| CN103605972B (en) | Non-restricted environment face verification method based on block depth neural network | |
| CN113221641B (en) | Video pedestrian re-identification method based on generation of antagonism network and attention mechanism | |
| CN113313164B (en) | Digital pathological image classification method and system based on super-pixel segmentation and graph convolution | |
| CN109033978B (en) | Error correction strategy-based CNN-SVM hybrid model gesture recognition method | |
| CN110414616B (en) | Remote sensing image dictionary learning and classifying method utilizing spatial relationship | |
| CN116448019B (en) | Intelligent detection device and method for quality flatness of building energy-saving engineering | |
| CN111695456A (en) | Low-resolution face recognition method based on active discriminability cross-domain alignment | |
| CN117237733A (en) | Breast cancer full-slice image classification method combining self-supervision and weak supervision learning | |
| CN115100709B (en) | Feature separation image face recognition and age estimation method | |
| CN113065520B (en) | Multi-mode data-oriented remote sensing image classification method | |
| CN111239137A (en) | Grain quality detection method based on transfer learning and adaptive deep convolution neural network | |
| CN114863151B (en) | Image dimension reduction clustering method based on fuzzy theory | |
| CN110991247B (en) | Electronic component identification method based on deep learning and NCA fusion | |
| Nishii et al. | Supervised image classification by contextual AdaBoost based on posteriors in neighborhoods | |
| CN113344069B (en) | Image classification method for unsupervised visual representation learning based on multi-dimensional relation alignment | |
| CN111259938A (en) | Manifold learning and gradient lifting model-based image multi-label classification method | |
| CN111274986B (en) | Dish identification and classification method based on image analysis | |
| CN117611830A (en) | Random class target positioning and counting method based on few sample labeling | |
| CN115131588B (en) | Image robust clustering method based on fuzzy clustering | |
| CN115272688A (en) | Small sample learning image classification method based on meta-features | |
| CN115439926A (en) | Small sample abnormal behavior identification method based on key region and scene depth | |
| Galea et al. | TCDetect: A new method of detecting the presence of tropical cyclones using deep learning | |
| CN112257787B (en) | Image semi-supervised classification method based on generation type dual-condition confrontation network structure |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |