Identity recognition method and system based on deep spectral clustering and computer storage medium

Abstract

The application discloses an identity recognition method and system based on deep spectral clustering and a computer storage medium, and relates to the technical field of identity recognition. The method comprises the following steps: obtaining identity data of a user; extracting features in the identity data by using a deep network model to obtain identity features; and clustering the identity features by using a spectral clustering model to obtain a category to which the identity features belong. The spectral clustering model is established based on a sparse KAN network. In the sparse KAN network, norm regularization and entropy regularization constraints are introduced, and an orthogonal constraint mechanism is also introduced in the last layer of the sparse KAN network. By introducing the norm regularization and entropy regularization constraints, the orthogonalization constraint and the manifold learning technology, the efficiency, stability and accuracy of identity recognition are effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of identity recognition, and particularly relates to an identity recognition method and system based on deep spectral clustering and a computer storage medium. BACKGROUND

[0002] Identity recognition is to recognize the category of a user, i.e., a legal user or an illegal user, through various data provided by the user, such as a face image, a fingerprint, and voiceprint data. At present, common identity recognition methods are mostly based on spectral clustering technology. Spectral clustering has been widely applied in various unsupervised learning tasks, such as identity recognition, due to its powerful global feature capturing capability and flexible affinity measurement mode. In recent years, the rise of deep learning has brought a new research paradigm to unsupervised clustering. According to the implementation process of a deep learning algorithm, it can be divided into data embedding based on a neural network and loss function calculation based on spectral clustering.

[0003] However, the existing identity recognition method based on deep spectral clustering has the following defects:

[0004] (1) The existing deep spectral clustering method adopts a fully connected layer to construct a network, which leads to an exponential increase in the number of parameters with the increase in network depth and input dimension, causing problems such as huge consumption of computing resources, high risk of overfitting, and directly leading to low processing efficiency of identity recognition.

[0005] (2) In the calculation of spectral clustering, the feature representation of data needs to be independent of each other to accurately reflect the similarity and difference between data points and to achieve effective clustering. However, in the calculation of the loss function using the spectral clustering structure, deep learning cannot guarantee the independence between the obtained network outputs, leading to poor stability of the identity recognition result.

[0006] (3) The traditional spectral clustering method relies on a static Gaussian kernel to construct an affinity matrix, which cannot be dynamically adjusted according to specific task requirements and is easily disturbed by noise in the original data, leading to poor accuracy of the identity recognition result. SUMMARY

[0007] The embodiments of the present application provide an identity recognition method, system and computer storage medium based on deep spectral clustering to solve the problems of low efficiency, poor stability and poor accuracy in identity recognition using the traditional spectral clustering technology in the prior art.

[0008] In one aspect, the embodiments of the present application provide an identity recognition method based on deep spectral clustering, comprising:

[0009] obtaining identity data of a user;

[0010] extracting features in the identity data by using a deep network model to obtain identity features;

[0011] The identity features are clustered by using a spectral clustering model to obtain a category to which the identity features belong, wherein the spectral clustering model is established based on a sparse KAN (Kolmogorov-Arnold network) network, norm regularization and entropy regularization constraints are introduced into the sparse KAN network, and an orthogonal constraint mechanism is also introduced into the last layer of the sparse KAN network; a loss function used by the sparse KAN network during training includes a regularization loss function and a spectral clustering loss function, the regularization loss function is established based on an activation function after the norm regularization and the entropy regularization constraints are introduced into the sparse KAN network, and the spectral clustering loss function is established based on an optimized affinity matrix; the optimization process of the affinity matrix includes: initializing the affinity matrix by using training samples, establishing an optimization objective function based on the initial affinity matrix, introducing a manifold learning technique into the optimization objective function, projecting the training samples into a low-dimensional space, and solving the optimization objective function to complete the optimization of the affinity matrix.

[0012] In another aspect, the embodiments of the present application also provide an identity recognition system based on deep spectral clustering, comprising:

[0013] a data acquisition module configured to acquire identity data of a user;

[0014] a deep network model configured to extract features in the identity data to obtain identity features;

[0015] a spectral clustering model configured to cluster the identity features to obtain a category to which the identity features belong.

[0016] The spectral clustering model is established based on a sparse KAN network, norm regularization and entropy regularization constraints are introduced into the sparse KAN network, and an orthogonal constraint mechanism is also introduced into the last layer of the sparse KAN network; a loss function used by the sparse KAN network during training includes a regularization loss function and a spectral clustering loss function, the regularization loss function is established based on an activation function after the norm regularization and the entropy regularization constraints are introduced into the sparse KAN network, and the spectral clustering loss function is established based on an optimized affinity matrix; the optimization process of the affinity matrix includes: initializing the affinity matrix by using training samples, establishing an optimization objective function based on the initial affinity matrix, introducing a manifold learning technique into the optimization objective function, projecting the training samples into a low-dimensional space, and solving the optimization objective function to complete the optimization of the affinity matrix.

[0017] In another aspect, the embodiments of the present application also provide a computer storage medium, which stores a plurality of computer instructions for causing a computer to execute the above method.

[0018] The identity recognition method, system and computer storage medium based on deep spectral clustering in the present application have the following advantages:

[0019] (1) By introducing norm regularization and entropy regularization constraints, the activation function matrix of the KAN network is sparsified, realizing the sparsity of the parameters within the layer and avoiding the excessive weight of the single-layer matrix. At the same time, the diversity of the activation function distribution is ensured through entropy regularization, and the regularization weight is dynamically adjusted during the training process. In the early stage of training, the constraint force is reduced to retain the features, and in the later stage, the constraint is enhanced to improve the sparsity, further reducing the computational complexity of the network, and thus improving the efficiency of identity recognition.

[0020] (2) An orthogonalization constraint layer is added to the last layer of the network, and Cholesky decomposition is used to impose an orthogonal constraint on the embedding matrix to ensure the independence and decoupling of the feature representation, avoid feature redundancy in the feature learning process, and solve the problems of unclear optimization path, slow convergence speed and gradient explosion caused by feature redundancy in traditional methods. The convergence speed of the model during the training process is improved, the discriminative ability of the features is enhanced, and thus the stability of the identity recognition result is improved.

[0021] (3) The manifold learning technique is used to generate weighted coefficient adjustment samples to balance the global and local features, and these weighted features are projected into a low-dimensional space to extract more distinctive state feature representations, avoid feature redundancy, promote more effective processing and analysis, solve the strong dependence of traditional spectral clustering on static Gaussian kernel, avoid the gradient instability and insufficient global structure capture caused by the kernel width hyperparameter, and improve the accuracy of the identity recognition result. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0023] Figure 1 The flowchart of the identity recognition method based on deep spectral clustering provided by the embodiments of the present application. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0025] Figure 1The flowchart of the identity recognition method based on deep spectral clustering provided by the embodiment of the present application. The embodiment of the present application provides an identity recognition method based on deep spectral clustering, comprising the following steps:

[0026] S100, obtaining identity data of a user.

[0027] Exemplarily, the identity data can be a face image, a fingerprint image or a voiceprint signal. After obtaining the identity data, it is necessary to perform data cleaning and preprocessing on the identity data. The data cleaning includes removing low-quality samples, repeated samples or incorrectly labeled data, and the preprocessing includes standardization and denoising and enhancement. The standardization is to unify the size, normalize the gray scale or perform mean-variance standardization on the data, and the denoising and enhancement is to expand the sample size by filtering, data enhancement and the like.

[0028] S110, extracting features in the identity data by using a deep network model to obtain identity features.

[0029] Exemplarily, for the image type identity data, the present application uses a CNN (Convolutional Neural Network) to extract identity features, and for the audio type identity data, the present application uses a Transformer to extract identity features.

[0030] S120, clustering the identity features by using a spectral clustering model to obtain a category to which the identity features belong, wherein the spectral clustering model is established based on a sparse KAN network, the norm regularization and entropy regularization constraints are introduced into the sparse KAN network, and the orthogonal constraint mechanism is also introduced into the last layer of the sparse KAN network. The loss function used by the sparse KAN network in training includes a regularization loss function and a spectral clustering loss function. The regularization loss function is established based on an activation function after the norm regularization and entropy regularization constraints are introduced into the sparse KAN network. The spectral clustering loss function is established based on an optimized affinity matrix. The optimization process of the affinity matrix includes: initializing the affinity matrix by using training samples, establishing an optimization objective function based on the initial affinity matrix, introducing a manifold learning technology into the optimization objective function, and projecting the training samples into a low-dimensional space to solve the optimization objective function to complete the optimization of the affinity matrix.

[0031] Exemplarily, the activation function matrix of the KAN network constructed by the present application is represented as:

[0032]

[0033] wherein, KAN ( x ) represents the result processed by the KAN network when the input training sample, i.e., the identity feature, is x , L is the number of layers of the activation function.l is the layer index of the activation function, is the layer activation function matrix, where each node has a different activation function, the Kolmogorov-Arnold representation theorem limits the number of hidden layers to 2 and the number of nodes in each hidden layer to , while the KAN network can have an arbitrary number of layers and each layer can contain an arbitrary number of nodes. The activation function can be represented as:

[0034]

[0035]

[0036] where, is the activation function in , is the layer index of the activation function, l is the layer index of the activation function, is the l layer activation function matrix, is the th activation function in the th layer, and are the weights of the two parts, respectively, is a linear unit function, and is composed of multiple spline basis functions. Typically, the number of node vectors in the spline function is 8.

[0037] Further, define the set of activation function matrices of each layer as , and impose the norm constraint in the following way:

[0038] Intra-layer sparsity constraint: impose a l 1 norm constraint on the activation function matrix of each layer to promote parameter sparsity within a specific layer and reduce invalid connections.

[0039] Global balance constraint: impose a l 2 norm constraint on the overall activation function matrix to avoid optimization imbalance caused by excessively large single-layer matrix weights.

[0040] The expression of the norm regularization is:

[0041]

[0042] where, denotes the norm regularization, L denotes the number of layers of the activation function, l is the layer index of the activation function, denotes the l layer activation function matrix, Indicates taking l 1-norm. This mechanism combines the advantages of both norms, achieving a balance between matrix sparsity and global structure.

[0043] To avoid the decrease in expressive power caused by sparsity, entropy regularization is introduced, which is expressed as:

[0044]

[0045] in, This represents entropy regularization. L This indicates the number of layers in the activation function. l The layer number is the activation function. Indicates the first l Layer ( i , j The normalized values ​​of the activation functions are calculated as follows:

[0046]

[0047] in, Indicates taking l 1-norm, l The layer number is the activation function. k Indicates the first The total number of nodes in the layer m For the first The total number of nodes in the layer.

[0048] By maximizing the entropy of the activation function distribution, matrix homogenization is prevented, ensuring the model's ability to represent features of high-dimensional data. The regularized loss function is expressed as:

[0049]

[0050] in, This represents the regularization loss function. t Indicates the current number of training iterations. This represents the total number of training sessions. Lambda This represents the regularization weight parameter. Norm regularization, This represents entropy regularization. The strategy reduces the regularization strength in the early stages of training to prevent premature feature pruning; later, it gradually strengthens the constraints to improve matrix sparsity.

[0051] Furthermore, this application introduces an orthogonal constraint mechanism into the sparse KAN network, optimizing the embedding space by enforcing the independence of feature vectors, thereby improving the convergence stability and feature discrimination ability of the clustering task. The mapping relationship of the sparse KAN network is defined. ,in As training samples, y To be compatible with input xcorresponding output. To ensure the strong independence of the feature vectors in the loss function, the orthogonal constraint is represented as:

[0052]

[0053] wherein, y denotes the output of the sparse KAN network when the input training sample is x , the upper index T denotes the transpose, denotes the expectation of , denotes the unit matrix, denotes the number of clusters. c

[0054] Through randomly selecting m training samples to form a small batch for iterative training, the orthogonal constraint is converted into a matrix form:

[0055]

[0056] wherein, m is the number of randomly selected training samples, is the embedding matrix, is stacked by rows, denotes the unit matrix, denotes the number of clusters. The orthogonalization calculation of the embedding matrix c is realized by Cholesky decomposition, Cholesky decomposition is performed on the target matrix to obtain a lower triangular matrix , which satisfies , the orthogonal matrix is reconstructed through the decomposition result, to ensure that strictly satisfies the unit matrix constraint, and to avoid optimization deviation caused by feature redundancy. In the embodiments of the present application, to ensure the real-time optimization of the loss function, randomly selects

[0057] samples as input each time of training iteration, dynamically updates the small batch data, enhances the adaptability of the model to the data distribution, and accelerates the convergence process of the orthogonal constraint. m

[0058] Further, after inputting the training sample, the similarity measure inside the training sample is dynamically learned, and the element of the affinity matrix , i.e., the affinity value is expected to be learned, so that the smaller distance between the th training sample and the th training sample corresponds to the larger affinity value . For the global optimal similarity expression of the sample, the initial optimization objective function can be represented as: ​​

[0059]

[0060] in, Represents vector standardization, Indicates the first The affinity vector of each training sample to other training samples. m This represents the number of randomly selected training samples. x i and x j They represent the first i and j training samples, Indicates taking l 2-norm, Affinity matrix The affinity value in This is a regularization parameter. To ensure the sparsity of the resulting affinity matrix, a regularization constraint is added to the traditional spectral clustering calculation to ensure... The optimal solution is exactly one of the following: A non-zero value.

[0061] Define training samples and Distance between We can obtain:

[0062]

[0063] in, d i Indicates training samples and The Euclidean distance between them. If we take d i By reordering the matrices from smallest to largest, we can ultimately obtain the globally optimal solution for the correlation matrix in the graph structure model, i.e., the affinity matrix. Initial value:

[0064]

[0065] in, Affinity matrix The affinity value in Indicates training samples and The distance between them c Indicates the number of clusters.

[0066] Considering the interference of noise in the training samples on the affinity matrix, a manifold learning technique is introduced to weight the training samples. Therefore, the initial optimization objective function can be updated as follows:

[0067]

[0068]

[0069] in, m This represents the number of randomly selected training samples. x i and x j They represent the first i and j training samples, Indicates taking l 2-norm, It is a diagonal matrix, and the diagonal elements of this diagonal matrix are... Used for feature weighting, balancing global and local features. Affinity matrix The affinity value in It is a regularization parameter. Represents vector standardization, Indicates the first The affinity vectors of each training sample to other training samples. By using manifold learning weighting, the influence of noise in the training samples is effectively reduced, and the ability of the affinity matrix to capture the local structure of the data is improved.

[0070] By introducing manifold learning techniques and projecting the training samples into a low-dimensional space, the aforementioned optimization objective function is updated again as follows:

[0071]

[0072]

[0073] in, m This represents the number of training samples randomly selected during the training process. for W transpose, Denotes the projection subspace matrix, satisfying the orthogonality constraint. Represents a diagonal matrix. x i and x j They represent the first i and j training samples, Indicates taking l 2-norm, s ij Representing the affinity matrix S ij The affinity value in and Represents the regularization parameter. Represents the trace of a matrix. Represents vector standardization, Indicates the first The affinity vector of each training sample to other training samples. diagonal matrix diagonal elements in I Represents the identity matrix. The purpose of this method is to prevent the projection matrix from degenerating into a zero matrix, ensuring the effectiveness of the feature mapping. Through this projection mechanism, more discriminative low-dimensional features are extracted, feature redundancy is avoided, and the accuracy of subsequent clustering tasks is improved.

[0074] To obtain the global optimum of the objective function, gradient descent is used to solve it. During the solution process, the objective function is first fixed... and s ij Solving for the optimized solution yields the following results. W Then fix W and s ij Solving for the optimized solution yields the following results. Finally based on and W Solving for the optimal s ij .

[0075] Specifically, first fix and s ij ,right The optimization objective function is:

[0076]

[0077] in, for W transpose, Denotes the projection subspace matrix. Represents the trace of a matrix. Represents a diagonal matrix. X Indicates input data, It is a lower triangular matrix. This represents the regularization parameter, calculated by... The former The eigenvectors corresponding to the smallest eigenvalues ​​are obtained, resulting in the optimized eigenvectors. .

[0078] Then fix and s ij ,right The objective function is:

[0079]

[0080]

[0081] wherein, is a diagonal matrix with the i th diagonal element, denotes the i th , is an intermediate matrix, denotes a regularization parameter, is a transpose of W , denotes a projection subspace matrix, denotes the i th , d denotes a data dimension of input data, X denotes input data, is a lower triangular matrix, denotes a dot product.

[0082] After optimization, the affinity matrix can be updated.

[0083] Further, the spectral clustering loss function is denoted as:

[0084]

[0085] wherein, denotes a spectral clustering loss function, y i and y j denote outputs of the sparse KAN network when the training sample is x i and x j respectively, denotes taking the l 2-norm, s ij is an affinity value in the optimized affinity matrix S ij .

[0086] Based on the regularization loss function and the spectral clustering loss function, the loss function used in the training of the sparse KAN network in the present application can be obtained as Loss :

[0087]

[0088] wherein, denotes a regularization loss function, denotes a spectral clustering loss function.

[0089] The embodiment of the present application also provides an identity recognition system based on deep spectral clustering, which comprises:

[0090] a data acquisition module, configured to acquire identity data of a user;

[0091] a deep network model, configured to extract features in the identity data to obtain identity features;

[0092] a spectral clustering model, configured to cluster the identity features to obtain a category to which the identity features belong;

[0093] The spectral clustering model is established based on a sparse KAN network, norm regularization and entropy regularization constraints are introduced into the sparse KAN network, and an orthogonal constraint mechanism is also introduced into the last layer of the sparse KAN network; a loss function used by the sparse KAN network in training comprises a regularization loss function and a spectral clustering loss function; the regularization loss function is established based on an activation function after the norm regularization and the entropy regularization constraints are introduced into the sparse KAN network; the spectral clustering loss function is established based on an optimized affinity matrix; the optimization process of the affinity matrix comprises: initializing the affinity matrix by using training samples, establishing an optimization objective function based on the initial affinity matrix, introducing a manifold learning technology into the optimization objective function, projecting the training samples into a low-dimensional space, and solving the optimization objective function to complete the optimization of the affinity matrix.

[0094] The embodiment of the present application also provides a computer storage medium, which stores a plurality of computer instructions, and the computer instructions are used for enabling a computer to execute the method.

[0095] Although the preferred embodiments of the present application have been described, those skilled in the art can make additional changes and modifications to the embodiments once they know the basic inventive concept. Therefore, the appended claims are intended to be interpreted as including all the changes and modifications falling within the scope of the present application.

[0096] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application is also intended to include these modifications and variations.

Claims

1. A method of identity recognition based on deep spectral clustering, characterized in that, The method comprises: obtaining identity data of a user, the identity data being a face image, a fingerprint image or a voiceprint signal; extracting features in the identity data by using a deep network model to obtain identity features; performing clustering on the identity features by using a spectral clustering model to obtain a category to which the identity features belong, wherein the spectral clustering model is established based on a sparse KAN network, norm regularization and entropy regularization constraints are introduced into the sparse KAN network, an orthogonal constraint mechanism is also introduced into a last layer of the sparse KAN network, a loss function used by the sparse KAN network during training comprises a regularization loss function and a spectral clustering loss function, the regularization loss function is established based on an activation function after the norm regularization and the entropy regularization constraints are introduced into the sparse KAN network, the spectral clustering loss function is established based on an optimized affinity matrix, and an optimization process of the affinity matrix comprises: initializing the affinity matrix by using training samples, establishing an optimization objective function based on an initial affinity matrix, introducing a manifold learning technique into the optimization objective function, projecting the training samples into a low-dimensional space, and solving the optimization objective function to complete optimization of the affinity matrix. The norm regularization constraint comprises: Intra-layer sparsity constraint: imposing l 1-norm constraint; global balance constraint: imposing l 2-norm constraint; An expression of the norm regularization is: wherein, denotes the norm regularization, L denotes the number of layers of activation functions, l is the layer index of the activation function, denotes the l activation function matrix composed of the denotes the l 1-norm; The entropy regularization is expressed as: wherein, denotes the entropy regularization, L denotes the number of layers of activation functions, l is the layer index of the activation function, denotes the normalized value of the l layer the i , j activation function; The regularization loss function is expressed as: wherein, denotes the regularization loss function, t denotes the current training number, denotes the total training number, The orthogonal constraint is expressed as: denotes the regularization weight parameter, denotes the norm regularization, denotes the entropy regularization. 2.The deep spectral clustering based identity recognition method of claim 1, wherein, The spectral clustering loss function is expressed as: wherein, y denotes the output of the sparse KAN network when the input training sample is x , the upper index T denotes the transpose, denotes the expectation of , the upper index denotes the identity matrix, c denotes the number of clusters; By randomly selecting m a small batch of training samples to perform iterative training, the orthogonal constraint is converted into a matrix form: wherein, m is the number of randomly selected training samples, is the embedding matrix, is the number of clusters, is stacked by row, is the identity matrix, c is the number of clusters. 3.The deep spectral clustering based identity recognition method of claim 1, wherein, After the manifold learning technique is introduced and the training samples are projected into the low-dimensional space, the optimization objective function is expressed as: wherein, represents the spectral clustering loss function, m represents the number of training samples randomly selected during the training process, y i and y j respectively represent the output of the sparse KAN network when the training sample is x i and x j respectively represent the output of the sparse KAN network when the training sample is represents taking l 2-norm, s ij is the affinity value in the optimized affinity matrix S ij . 4.The deep spectral clustering based identity recognition method of claim 1, wherein, The method comprises: wherein, m denotes the number of randomly selected training samples in the training process, is the transpose of W denotes the projection subspace matrix, denotes the diagonal matrix, x i and x j denote the i-th and j-th training samples, i and j denote the i-th and j-th training samples, denotes the 2-norm, l s ij denotes the affinity value in the optimized affinity matrix S ij and denote the regularization parameters, denotes the trace of a matrix.​​​ 5. The method of claim 4, wherein, Solving the optimization objective function using gradient descent method, in the solving process, first fix and s ij , the solving obtains the optimized W , then fix W and s ij , the solving obtains the optimized , finally, based on and W solving the optimal s ij .

6. System for applying the method for identity recognition based on deep spectral clustering according to any of claims 1 to 5, characterized in that, a data acquisition module configured to acquire identity data of a user; a deep network model configured to extract features in the identity data to obtain identity features; a spectral clustering model configured to perform clustering on the identity features to obtain a category to which the identity features belong; the spectral clustering model is established based on a sparse KAN network, norm regularization and entropy regularization constraints are introduced into the sparse KAN network, an orthogonal constraint mechanism is also introduced into a last layer of the sparse KAN network, a loss function used by the sparse KAN network during training comprises a regularization loss function and a spectral clustering loss function, the regularization loss function is established based on an activation function after the norm regularization and the entropy regularization constraints are introduced into the sparse KAN network, the spectral clustering loss function is established based on an optimized affinity matrix, and an optimization process of the affinity matrix comprises: initializing the affinity matrix by using training samples, establishing an optimization objective function based on an initial affinity matrix, introducing a manifold learning technique into the optimization objective function, projecting the training samples into a low-dimensional space, and solving the optimization objective function to complete optimization of the affinity matrix. The computer storage medium stores a plurality of computer instructions for causing a computer to execute the method of any one of claims 1-5.

7. A computer storage medium, characterized in that The computer storage medium stores a plurality of computer instructions for causing a computer to execute the method of any one of claims 1-5.

Images