Encrypted image clustering method and encrypted image clustering system based on logistic mapping
By using Logistic mapping function on the user side for image scrambling and encryption, combined with feature extraction and clustering analysis of cloud servers, the problem of low efficiency of image data encryption and clustering in the existing technology is solved, and the efficient and security of image data is achieved Transmission and accurate clustering.
Patent Information
- Application Number
- CN202410979587.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-12
AI Technical Summary
The existing image data encryption and data integrity schemes are complex and consume a lot of computing power, so they cannot accurately detect data tampering nodes. The traditional encryption method is slow to encrypt and poor clustering, and cannot meet the needs of image privacy protection and efficient clustering at the same time.
The encrypted image clustering method based on Logistic mapping is adopted to scramble the image by the user side, and then feature extraction and clustering analysis are performed on the cloud server to ensure the confidentiality and clustering effect of the image data.
It improves the randomness and security of encryption, keeps image features unchanged, enhances image security and clustering accuracy, is suitable for large-scale image data processing, and realizes the secure transmission and effective clustering of image data.
Smart Images

Figure CN118941830A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of information security, and in particular to an encrypted image clustering method and an encrypted image clustering system based on logistic mapping. Background Art
[0002] With the rapid development of big data and information technology, the acquisition, transmission and storage of image data have become increasingly common. The widespread application of image data has put forward urgent demands for data security and privacy, and with it comes increasing attention to image privacy security, especially image data involving personal privacy or commercial secrets. Data integrity is also an important attribute of many software systems, and it is especially important in cloud computing. In today's era of huge demand for cloud computing platforms, users upload local data to cloud servers to save local storage resources. However, users may want to monitor their data and detect whether unauthorized attackers have tampered with the database and whether the cloud server has lost the original data.
[0003] The current image data encryption and data integrity solutions have the following problems:
[0004] 1. The solution is too complicated and consumes the computing power of the cloud server to calculate the verification data. During the integrity verification process, when the cloud server feeds back the data required for verification to the local computer, the communication volume is relatively large. It is impossible to accurately detect which specific node has a problem.
[0005] 2. Traditional image encryption methods usually use symmetric key or public key encryption algorithms, such as AES or RSA. Although these algorithms can ensure the confidentiality of image data to a certain extent, they often have problems such as slow encryption speed and poor clustering effect, which is particularly prominent when processing large amounts of image data. Therefore, existing image encryption and image clustering methods often cannot meet the needs of image privacy protection and efficient clustering at the same time.
[0006] To solve the above problems, a new image processing solution is needed that can take into account both the privacy and security of images and the clustering effect. It can ensure the confidentiality of image data and ensure that it is not easily stolen or tampered with during transmission. It can also ensure the effectiveness of the clustering effect of the encrypted image and will not affect the clustering analysis of the image data. These two aspects are the focus of encrypted image clustering and are of great significance to the security and analyzability of image data. Summary of the invention
[0007] To achieve the above objectives, the present invention designs a model for encrypted image clustering, which encrypts images while ensuring security and preventing malicious tampering, while retaining the characteristics of the images, thereby achieving secure transmission of the images. At the same time, the clustering speed and accuracy are improved.
[0008] The invention provides an encrypted image clustering method and an encrypted image clustering system based on logistic mapping.
[0009] A method for clustering encrypted images based on logistic mapping, the system used includes a user end and a cloud server, the two are connected in communication, and the method includes the following steps:
[0010] Step 1: The user selects the image to be encrypted, uses the logistic mapping function to perform sub-block scrambling and pixel scrambling, and encrypts the image data;
[0011] Step 2: The user uploads the encrypted image data;
[0012] Step 3: The cloud server extracts features from the encrypted image;
[0013] Step 4: Repeat the above steps until the user has uploaded all encrypted images to the cloud server, and the cloud server has completed the feature extraction of all images;
[0014] Step 5: The cloud server performs cluster analysis on the image based on the extracted features;
[0015] Step 6: The client sends the clustering result to the client.
[0016] Furthermore, step one includes the following steps:
[0017] Step 1.1: The user end divides the original image into N sub-blocks, each sub-block contains M×M pixels;
[0018] Step 1.2: The user terminal provides a key to generate an initial seed x0 and the parameter λ of the logistic mapping function, and then converts x0 into a floating-point number in the range [0,1]. The logistic mapping function is used to generate a pseudo-random floating-point number sequence, and then the pseudo-random integer sequences P1 and P2 are generated according to the size relationship of the floating-point numbers.
[0019] P1 is an integer sequence containing n values, P1={P 1,1 ,P 1,2 ,…,P 1,i ,…P 1,n}, n = N; P2 is an integer sequence containing m values, P2 = {P 2,1 ,P 2,2 ,…,P 2,j …,O 2,m}, m = M × M;
[0020] Step 1.3: The user end uses pseudo-random integer sequences P1 and P2 to scramble sub-blocks and pixels within blocks respectively.
[0021] Furthermore, in the sub-block scrambling process in step 1.3, for the i-th sub-block, the permutation is performed by remapping the original pixel position to the new position:
[0022] Index i =(p 1,i +i)mod(N)
[0023] Where i is the number of the i-th sub-tile, Index i is the new position of the ith sub-block of the original image, p 1,i Corresponding to the i-th number in the pseudo-random integer sequence P1.
[0024] Furthermore, the process of pixel scrambling within a block in step 1.3 is as follows: for the i-th sub-block, use the pseudo-random sequence P2 to perform pixel scrambling:
[0025]
[0026] Where NPixel i,j is the new position of the jth pixel in the ith subtile, and OPixel i,j represents the original pixel at the jth position in the i-th sub-block, p 2,j +j is the shift operation on the j-th pixel.
[0027] Furthermore, step three is specifically as follows:
[0028] Step 3.1: Count the total number of pixels in the encrypted image;
[0029] Step 3.2: Traverse each pixel in the encrypted image to obtain its RGB color value (R x,y ,G x,y ,B x,y ), and then use the following calculation to convert these values to grayscale values Gray x,y ;
[0030] Gray x,y =0.299·R x,y +0.587 G x,y +0.114 B x,y
[0031] Set all gray values to Gray z,t Stored in the grayscale array Gray; x, y represent the horizontal and vertical coordinate values of the pixel respectively;
[0032] Step 3.3: Calculate sub-block B i The average gray value within is calculated as follows:
[0033]
[0034] Average i is the i-th sub-block B i The average gray value within Represents sub-block B i The gray value of the pixel with the internal coordinate value (x, y);
[0035] Step 3.4: The average grayscale value of each sub-block constitutes the feature array corresponding to the encrypted image.
[0036] Furthermore, in step five, the cloud server performs clustering according to the feature array corresponding to each image.
[0037] An encrypted image clustering system for implementing the aforementioned encrypted image clustering method based on logistic mapping includes a user terminal and a cloud server, wherein the user terminal is in communication connection with the cloud server, the user terminal includes an encryption module and a first signal transmission module; the cloud server includes an image feature extraction module, a cluster analysis module, and a second signal transmission module;
[0038] The encryption module is based on the logistic mapping function and uses sub-block scrambling and pixel scrambling to encrypt image data;
[0039] The first signal transmission module and the second signal transmission module are used for communication between the cloud server and the user end;
[0040] The image feature extraction module processes the pixel values of the encrypted image, counts the number of sub-blocks in the image, and extracts features from the pixel points of the image and stores them as a feature array;
[0041] The clustering analysis module is used to perform clustering according to the feature arrays corresponding to each encrypted image.
[0042] Furthermore, the encryption process of the encryption module is:
[0043] Step 1.1: The user end divides the original image into N sub-blocks, each block contains M×M pixels;
[0044] Step 1.2: The user provides a key to generate the initial seed x0 and the parameter λ of the logistic mapping function, and then converts x0 into a floating point number in the range [0,1]. n+1 =λx n (1-x n ) generates a pseudo-random floating point number sequence, and then generates pseudo-random integer sequences P1 and P2 according to the magnitude relationship of the floating point numbers;
[0045] P1 is an integer sequence containing n values, P1={P 1,1 ,P1,2 ,…,P 1,i ,…P 1,n}, n = N, P2 is an integer sequence containing m values, P2 = {P 2,1 ,P 2,2 ,…,P 2,j …,P 2,m}, m = M × M;
[0046] Step 1.3: The user end uses pseudo-random integer sequences P1 and P2 to scramble sub-blocks and pixels within blocks respectively.
[0047] Furthermore, the sub-block scrambling is specifically as follows: for the i-th sub-block, the arrangement is performed by remapping the original pixel position to the new position:
[0048] Index i =(p 1,i +i)mod(N)
[0049] Where i represents the number of the i-th sub-tile, Index i is the new position of the ith sub-block of the original image, p 1,i Corresponding to the i-th number in the pseudo-random integer sequence P1.
[0050] Furthermore, the pixel scrambling within the block is specifically as follows: for each sub-block B i , use the pseudo-random integer sequence P2 to scramble the pixels within the block:
[0051]
[0052] Where NPixel i,j is the new position of the jth pixel in the ith subtile in the original image, and OPixel i,j represents the original pixel at the jth position in the i-th sub-block, p 2,j +j is the shift operation on the j-th pixel.
[0053] The technical effects of the present invention are as follows:
[0054] 1. The present invention adopts the random number sequence generated by Logistic mapping as the key of the encryption algorithm, which improves the randomness and security of encryption. At the same time, it utilizes the characteristics of Logistic mapping to keep the image features unchanged and improve the accuracy of image clustering.
[0055] 2. The present invention can successfully encrypt images while maintaining the integrity of image features, enhancing the security of images without changing their content.
[0056] 4. The present invention can be flexibly applied to image data of different contents, has good scalability, and is suitable for processing and encrypting large-scale image data.
[0057] 5. By using the present invention, local users can encrypt image data, effectively protecting the security and privacy of image data and preventing unauthorized access and theft. At the same time, the features of the encrypted image remain unchanged, improving the effect and accuracy of clustering. It can be combined with an efficient clustering algorithm to achieve accurate clustering of encrypted images, which can effectively meet the needs of image data security and clustering analysis in the digital age. BRIEF DESCRIPTION OF THE DRAWINGS
[0058] Figure 1 This is a system model diagram used in the present invention. DETAILED DESCRIPTION
[0059] The present invention proposes a safe, accurate and reliable encrypted image clustering method and a system based on the method. The aforementioned method is an encrypted image clustering based on Logistic mapping to solve the problems of low efficiency of traditional image encryption methods and information leakage in the image clustering process. The entire encryption process generates a pseudo-random integer sequence based on Logistic mapping to scramble the image, thereby enhancing the security of the image without changing its content. The present invention can not only effectively protect the privacy security of the image, but also realize the clustering of the encrypted image.
[0060] like Figure 1 As shown, the system used in the present invention includes two participants: a user end and a semi-trusted cloud server. The user end has absolute ownership of the data, encrypts and uploads the data, and the encrypted data is stored in the cloud server for cluster analysis.
[0061] The user end includes an encryption module and a first signal transmission module; the cloud server includes an image feature extraction module, a cluster analysis module, and a second signal transmission module.
[0062] The encryption module is based on the logistic mapping function and uses sub-block scrambling and pixel scrambling to encrypt image data;
[0063] The first signal transmission module and the second signal transmission module are used for communication between the cloud server and the user end;
[0064] The image feature extraction module processes the pixel values of the encrypted image, counts the number of sub-blocks in the image, and extracts features from the pixel points of the image and stores them as a feature array;
[0065] The clustering analysis module is used to perform clustering according to the feature arrays corresponding to each encrypted image.
[0066] The system's workflow is:
[0067] Step 1: The user end selects the image to be encrypted and encrypts the image data. This step is completed by the encryption module. In this step, the user end selects a key, generates initial parameters, and uses the logistic mapping function to generate a pseudo-random floating-point number sequence. Then, according to the size relationship of the floating-point numbers, an integer sequence P1 and P2 are generated to scramble the image blocks and the pixels in the blocks, and change the positions of the image blocks and pixels in each block.
[0068] The specific method for encrypting image data is:
[0069] Step 1.1: The user end divides the original image into N sub-blocks, each of which contains M×M pixels. The value of N can be determined as needed, and once N is determined, M is naturally determined.
[0070] Step 1.2: The user provides a key, which is a string or number of any length, and uses a hash function to process the key to generate an initial seed x0 and a parameter λ of the mapping function. The initial seed x0 is then converted to a floating point number in the range [0,1] and subsequently used as the input parameter x in the logistic mapping function. n The value range of λ is 3.5699456 to 4, and the specific value is determined according to the requirements.
[0071] Through the above process, the key is converted into an initial seed. The initial seed x0 and the parameter λ of the logistic mapping function are used to initialize the logistic mapping function. The logistic mapping function generates a pseudo-random floating point number sequence, and then generates a pseudo-random integer sequence P1 and P2 according to the size relationship of the floating point numbers. Such a conversion process ensures the security and randomness of the key, while providing sufficient complexity to enhance the security of image encryption. P1 is used for image block scrambling and pixel scrambling within a block.
[0072] P1 is an integer sequence containing n values, P1={P 1,1 ,P 1,2 ,…,P 1,i ,…P 1,n}, n=N.
[0073] P2 is an integer sequence containing m values, P2={P 2,1 ,P 2,2 ,…,P 2,j …,P 2,m},m=M×M.
[0074] Step 1.3: The user end uses pseudo-random integer sequences P1 and P2 to scramble sub-blocks and pixels within sub-blocks respectively.
[0075] Sub-tile scrambling involves rearranging the positions within individual sub-tiles of an image, thereby increasing the complexity of encryption. The original image is divided into different sub-tiles, each of which undergoes a permutation process based on a pseudo-random integer sequence generated using a logistic mapping function.
[0076] Pixel scrambling further disrupts the order of pixels within each sub-block, which helps in image encryption. This step involves applying another pseudo-random integer sequence generated using a logical mapping function to each sub-block, thereby increasing the level of encryption security. The combination of sub-block scrambling and pixel scrambling techniques effectively protects the confidentiality and integrity of digital images, preventing unauthorized access or tampering. The following is a detailed description of the process:
[0077] The principle of sub-block scrambling is: the original image I is divided into N sub-blocks, each sub-block contains M×M pixels. The value of N can be determined as needed. Using the logistic mapping function x n+1 =λx n (1-x n ) generates a pseudo-random floating-point number sequence containing n values, and then generates a pseudo-random integer sequence P1 containing n values according to the size of the value in this pseudo-random floating-point number sequence, which is used to scramble the original image. i , the permutation is performed by remapping the original pixel positions to new positions:
[0078] Index i =(p 1,i +i)mod(N)
[0079] Where i represents the number of the i-th sub-tile. i is the new position of the ith sub-block of the original image, p 1,i Corresponding to the i-th number in the pseudo-random integer sequence P1.
[0080] The process of pixel scrambling is as follows: During the encryption process, the logistic mapping function input parameter x n Another floating point number converted from the initial seed. i , use the pseudo-random integer sequence P2 to scramble pixels:
[0081]
[0082] Where NPixel i,j is the new position of the jth pixel in the ith subtile in the original image, and OPixel i,j represents the original pixel at the jth position in the i-th sub-block, p 2,j +j is the shift operation on the j-th pixel.
[0083] Step 2: The user end uploads data. The user end sends the encrypted image data to the cloud server, and preferably stores the encrypted image at the same time to facilitate subsequent processing operations. This step is completed by the first signal transmission module.
[0084] Step 3: Extract encrypted image features. The cloud server processes the pixel values of the encrypted image, counts the number of sub-blocks in the image, and extracts features from the pixels of the image and stores them as feature arrays. This feature extraction ensures that the features of the image remain unchanged even after encryption. This step is completed by the image feature extraction module.
[0085] Feature extraction is a basic step in image analysis, which extracts relevant information from raw pixel data to represent the image content in a more abstract and meaningful way. The process of image feature extraction is as follows:
[0086] Step 3.1: Count the total number of pixels in the encrypted image, denoted as N pixels .
[0087] Step 3.2: Traverse each pixel in the encrypted image to obtain its RGB color value (R x,y ,G x,y ,B x,y ). Then use the following calculation to convert these values to grayscale values Gray x,y :
[0088] Gray x,y =0.299·R x,y +0.587 G x,y +0.114 B x,y
[0089] Set all gray values to Gray x,y Stored in the grayscale array Gray. x, y represent the horizontal and vertical coordinate values of the pixel respectively.
[0090] Step 3.3: Calculate sub-block B i The average gray value within is calculated as follows:
[0091]
[0092] Average i is the average grayscale value in the i-th sub-block, Represents the grayscale value of the pixel with coordinates (x, y) in the i-th sub-block.
[0093] Step 3.4: The average grayscale value of each sub-block constitutes the feature array corresponding to the encrypted image.
[0094] The feature extraction process involves only the statistical calculation of pixel values in the image, without considering pixel position or order. Therefore, even after encryption, the characteristics of the image remain intact, ensuring the consistency of recognition and analysis of images before and after encryption.
[0095] Step 4: Repeat the above steps until the user has uploaded all encrypted images to the cloud server, and the cloud server has completed the feature extraction of all images.
[0096] Step 5: Image clustering analysis. This step is completed by the clustering analysis module. The cloud server clusters each image according to its corresponding feature array, and selects the appropriate clustering method according to the data. For example, for color histogram data, K-means or hierarchical clustering can be used; for texture feature data, K-means, hierarchical clustering or density clustering (DBSCAN) can be used; for shape feature data, K-means or spectral clustering can be used. The optimization goal is to select the most suitable clustering method according to the different image features to obtain accurate and efficient clustering results. Finally, the clustering results are returned to the user end, so that the user end can analyze the results and make decisions.
[0097] Image clustering is the process of grouping images with similar features into the same category. The goal of image clustering is to group encrypted image data so that images in the same group are similar to each other, while images in different groups are very different.
[0098] According to the clustering goal, the clustering process is optimized. The algorithm starts by randomly selecting data points as the initial cluster centers, and then enters the iterative process. For each data point, the Euclidean distance to each cluster center is calculated, and the data point is assigned to the cluster to which the nearest cluster center belongs. Then, the new center of each cluster is calculated, that is, the mean of all data points in the cluster. These two steps are iterated continuously, updating the distribution of data points and cluster centers until the change in cluster centers is less than the preset threshold or the predetermined maximum number of iterations is reached. Through this process, the algorithm gradually minimizes the sum of squares within the cluster, that is, the sum of the squares of the distances from each data point to the cluster center to which it belongs, thereby optimizing the clustering results.
[0099] Step 6: Send clustering results. The cloud server sends the final clustering results to the user end, providing a basis for data mining and intelligent analysis for the user end, and providing an important reference for decision-making. This step is completed by the second signal transmission module.
Claims
1. An encrypted image clustering method based on logistic mapping, characterized in that: The system used includes a user terminal and a cloud server, which are connected in communication. The method includes the following steps: Step 1: The user selects the image to be encrypted, uses the logistic mapping function to perform sub-block scrambling and pixel scrambling, and encrypts the image data; Step 2: The user uploads the encrypted image data; Step 3: The cloud server extracts features from the encrypted image; Step 4: Repeat the above steps until the user has uploaded all encrypted images to the cloud server, and the cloud server has completed the feature extraction of all images; Step 5: The cloud server performs cluster analysis on the image based on the extracted features; Step 6: The client sends the clustering result to the client.
2. The encrypted image clustering method based on logistic mapping as claimed in claim 1, characterized in that: Step 1 includes the following steps: Step 1.1: The user end divides the original image into N sub-blocks, each sub-block contains M×M pixels; Step 1.2: The user terminal provides a key to generate an initial seed x0 and the parameter λ of the logistic mapping function, and then converts x0 into a floating-point number in the range [0,1]. The logistic mapping function is used to generate a pseudo-random floating-point number sequence, and then the pseudo-random integer sequences P1 and P2 are generated according to the size relationship of the floating-point numbers. P1 is an integer sequence containing n values, P1={P 1,1 ,P 1,2 ,…,P 1,i ,…P 1,n }, n = N, P2 is an integer sequence containing m values, P2 = {P 2,1 ,P 2,2 ,…,P 2,j …,P 2,m }, m = M × M; Step 1.3: The user end uses pseudo-random integer sequences P1 and P2 to scramble sub-blocks and pixels within blocks respectively.
3. The encrypted image clustering method based on logistic mapping as claimed in claim 2, characterized in that: In the sub-tile scrambling process in step 1.3, for the i-th sub-tile, the permutation is performed by remapping the original pixel positions to the new positions: Index i =(p 1,i +i)mod(N) Where i is the number of the i-th sub-tile, Index i is the new position of the ith sub-block of the original image, p 1,i Corresponding to the i-th number in the pseudo-random integer sequence P1.
4. The encrypted image clustering method based on logistic mapping as claimed in claim 2 or 3, characterized in that: The process of pixel scrambling within a block in step 1.3 is as follows: for the i-th sub-block, use the pseudo-random sequence P2 to scramble the pixels: Where NPixel i,j is the new position of the jth pixel in the ith subtile, and OPixel i,j represents the original pixel at the jth position in the i-th sub-block, p 2,j +j is the shift operation on the j-th pixel.
5. The encrypted image clustering method based on logistic mapping as claimed in claim 2 or 3, characterized in that: Step three is as follows: Step 3.1: Count the total number of pixels in the encrypted image; Step 3.2: Traverse each pixel in the encrypted image to obtain its RGB color value (R x,y ,G x,y ,B x,y ), and then use the following calculation to convert these values to grayscale values Gray x,y ; Gray x,y =0.299·R x,y +0.587·G x,y +0.114·B x,y Set all gray values to Gray x,y Stored in the grayscale array Gray; x, y represent the horizontal and vertical coordinate values of the pixel respectively; Step 3.3: Calculate sub-block B i The average gray value within is calculated as follows: Average i is the i-th sub-block B i The average gray value within Represents sub-block B i The gray value of the pixel with the internal coordinate value (x, y); Step 3.4: The average grayscale value of each sub-block constitutes the feature array corresponding to the encrypted image.
6. The encrypted image clustering method based on logistic mapping as claimed in claim 5, characterized in that: In step five, the cloud server performs clustering based on the feature array corresponding to each image.
7. An encrypted image clustering system for implementing the encrypted image clustering method based on logistic mapping as claimed in any one of claims 1 to 6, characterized in that: The system comprises a user terminal and a cloud server, wherein the user terminal is in communication connection with the cloud server, the user terminal comprises an encryption module and a first signal transmission module; the cloud server comprises an image feature extraction module, a cluster analysis module, and a second signal transmission module; The encryption module is based on the logistic mapping function and uses sub-block scrambling and pixel scrambling to encrypt image data; The first signal transmission module and the second signal transmission module are used for communication between the cloud server and the user end; The image feature extraction module processes the pixel values of the encrypted image, counts the number of sub-blocks in the image, and extracts features from the pixel points of the image and stores them as a feature array; The clustering analysis module is used to perform clustering according to the feature arrays corresponding to each encrypted image.
8. The encrypted image clustering method according to claim 7, characterized in that: The encryption process of the encryption module is: Step 1.1: The user end divides the original image into N sub-blocks, each block contains M×M pixels; Step 1.2: The user provides a key to generate the initial seed x0 and the parameter λ of the logistic mapping function, and then converts x0 into a floating point number in the range [0,1]. n+1 =λx n (1-x n ) generates a pseudo-random floating point number sequence, and then generates pseudo-random integer sequences P1 and P2 according to the magnitude relationship of the floating point numbers; P1 is an integer sequence containing n values, P1={P 1,1 ,P 1,2 ,…,P 1,i ,…P 1,n }, n = N, P2 is an integer sequence containing m values, P2 = {P 2,1 ,P 2,2 ,…,P 2,j …,P 2,n }, m = M × M; Step 1.3: The user end uses pseudo-random integer sequences P1 and P2 to scramble sub-blocks and pixels within blocks respectively.
9. The encrypted image clustering method according to claim 8, characterized in that: Sub-block scrambling is performed as follows: For the i-th sub-block, the permutation is performed by remapping the original pixel positions to the new positions: Index i =(p 1,i +i)mod(MN Where i represents the number of the i-th sub-tile, Index i is the new position of the ith sub-block of the original image, p 1,i Corresponding to the i-th number in the pseudo-random integer sequence P1.
10. The encrypted image clustering method based on logistic mapping according to claim 8, characterized in that: The specific pixel scrambling within the block is: for each sub-block B i , use the pseudo-random integer sequence P2 to scramble the pixels within the block: Where NPixel i,j is the new position of the jth pixel in the ith subtile in the original image, and OPixel i,j represents the original pixel at the jth position in the i-th sub-block, p 2,j +f is the shift operation on the j-th pixel.
Citation Information
Patent Citations
Digital image encrypting and decrypting method
CN102609893A
Oversampling-correction-based large-scale image clustering method for cloud computing environment
CN106250907A
Searchable image encryption algorithm
CN106875325A
Encrypted face recognition method based on ciphertext domain
CN107045627A
Heterogeneous rock digital rock core modeling method based on K-means clustering algorithm
CN113515847A
