132 results about "K-SVD" patented technology

A signal processing system adapted for sparse representation of signals is provided, comprising: (i) one or more training signals; (ii) a dictionary containing signal-atoms; (iii) a representation of each training signal using a linear combination of said dictionary's signal-atoms; (iv) means for updating the representation of the training signal; (v) means for updating the dictionary one group of atoms at a time, wherein each atom update may include all representations referring to said updated atom; and (vi) means for iterating (iv) and (v) until a stopping rule is fulfilled. The system uses the K-SVD algorithm for designing dictionaries for sparse representation of signals.

The invention discloses a SAR (Synthetic Aperture Radar) image segmentation technique based on dictionary learning and sparse representation, and mainly solves the problems that the existing feature extraction needs a lot of time and some defects exist in the distance measurement. The method comprises the following steps: (1) inputting an image to be segmented, and determining a segmentation class number k; (2) extracting a p*p window for each pixel point of the image to be segmented so as to obtain a test sample set, and randomly selecting a small amount of samples from the test sample set to obtain a training sample set; (3) extracting wavelet features of the training sample set; (4) dividing the training sample set by using a spectral clustering algorithm; (5) training a dictionary by using a K-SVD (Kernel Singular Value Decomposition) algorithm for each class of training samples; (6) solving sparse representation vectors of the test sample on the dictionary; (7) calculating a reconstructed error function of the test sample; and (8) calculating a test sample label according to the reconstructed error function to obtain the image segmentation result. The invention has the advantages of high segmentation speed and favorable effect; and the technique can be further used for automatic target identification of SAR images.

The invention discloses an image super-resolution reconstruction method based on self-similarity and structural information constraint. The image super-resolution reconstruction method based on the self-similarity and the structural information constraint comprises the achieving steps: (1) taking z images from an image base, carrying out imitating quality degradation on each image, generating a low-resolution image, and constructing a dictionary training sample set; (2) in the dictionary training sample set, learning a pair of high resolution ratio dictionary and low resolution ratio dictionary through a kernel singular value decomposition (K-SVD) method; (3) for a to-be-processed low-resolution image Xt, with scale rotation transform utilized, searching k similar blocks {p1,p2,...,pk} which are mostly similar with an image block xi; (4) carrying out constraint solution on the image block xi through the obtained k similar blocks to obtain a sparse presentation coefficient A; (5) obtaining k reconstruction results through the sparse presentation coefficient A combined with a high-resolution dictionary DH; (6) utilizing a low rank presentation model, amending a similarity degree of the reconstruction results with the similar blocks {p1,p2,...,pk} under the low resolution utilized; (7) obtaining a final result through the amended similarity degree combined with the reconstruction results; and repeating the steps in sequence and obtaining a final high-resolution image YH. The image super-resolution reconstruction method based on the self-similarity and the structural information constraint has the advantages that structural information of the reconstruction results keeps good, and the image super-resolution reconstruction method can be used for image recognition and target classification.

The invention discloses an image super-resolution reconstruction method based on dictionary learning and structure similarity, mainly solving the problem that a reconstructed image based on the prior art has a fuzzy surface and a serious marginal sawtooth phenomenon. The image super-resolution reconstruction method comprises the following implementation steps of: (1) acquiring a training sample pair; (2) learning a pair of high/low-resolution dictionaries by using structural similarity (SSIM) and K-SVD (K-Singular Value Decomposition) methods; (3) working out a sparse expression coefficient of an input low-resolution image block; (4) reestablishing a high-resolution image block Xi by using the high-resolution dictionaries and the sparse coefficient; (5) fusing the high-resolution image block Xi to obtain a high-resolution image X'I subjected to information fusion; (6) obtaining a high-resolution image X according to the high-resolution image X'I; and (7) carrying out high-frequency information enhancement on the high-resolution image X through error compensation to obtain a high-resolution image subjected to high-frequency information enhancement. A simulation experiment shows that the image super-resolution reconstruction method has the advantages of clear image surface and sharpened margin and can be used for image identification and target classification.

The invention provides a synthetic aperture radar (SAR) image bionic recognition method based on sample generation and nuclear local feature fusion and belongs to the field of image processing technologies and SAR target recognition. According to the method, a super complete training sample set is firstly constructed for training to obtain geometry manifold, then a sample to be recognized is recognized, specifically, each sample is firstly subjected to image denoising by a K-SVD dictionary learning method, and object region extraction is achieved by means of an object centroid method; and feature extraction is performed respectively by combining local phase quantization (LPQ) and a Gabor filtering method, feature fusion is performed, finally, classification is performed by covering of high-dimensional geometry manifold, and recognition is performed by a bionic mode. According to the SAR image bionic recognition method based on sample generation and nuclear local feature fusion, inhibiting effects of image coherent noises are obvious, SAR image features can be effectively extracted, the problem of the unstable extracted features, which is caused by changes of attitude angles of SAR images, is solved, the recognition accuracy is high, and the method has good robustness.

The invention relates to a residual-based ultra-resolution image reconstruction method, which specifically comprises the following steps of: first calculating residuals between original high-resolution images and images obtained by performing interpolation amplification on low-resolution images; then establishing sample pairs by using the characteristics of low-resolution image samples and corresponding image residuals, classifying the sample pairs by taking the low-resolution image samples as references and adopting K-averaging, and training each type of sample pair by adopting a K-singular value decomposition (K-SVD) method to obtain dictionary pairs of the low-resolution image samples and the image residuals; and finally selecting a dictionary pair according to a Euclidean distance between a test sample and a type center, calculating the weighted sum of image residuals reconstructed by each type with similar Euclidean distances with the test sample as a final reconstructed image residual, and obtaining a high-resolution image by combining interpolation results of the low-resolution images. Only the image residuals are required to be reconstructed, and the high-resolution image can be reconstructed by combining the interpolated images, so that edge detail reconstruction results of the high-resolution image are improved.

The invention discloses a behavior identification method based on multi feature fusion. The behavior identification method comprises the steps that S1, preprocessing is carried out on videos of a training set and a test set; S2, feature point detection is carried out on the videos which are preprocessed in the step of S1; S3, descriptors which represent different information is extracted from feature cubes in the videos, so as to form comprehensive descriptors; S4, the descriptors, which are extracted in the step of S3, of the training set are used, and a K-SVD algorithm is used to train a dictionary; and S5, feature fusion is carried out on the descriptors, which are extracted in the step of S3, of the test set, and classification is carried out through a cascade dictionary classification algorithm. According to the invention, multi feature fusion is carried out on energy information, spatial information and time information; essential motion features are abstracted; comprehensive descriptor information extracting is realized; and the system robustness is great.

The invention discloses an image super-resolution reconstruction method based on dictionary learning and structure clustering, mainly solving the problem that a reconstructed image based on the prior art has a fuzzy surface and a serious marginal sawtooth phenomenon. The image super-resolution reconstruction method comprises the following implementation steps of: (1) acquiring training samples; (2) structurally clustering the training samples; (3) training by using OMP (Orthogonal Matching Pursuit) and K-SVD (K-Singular Value Decomposition) methods to obtain various dictionaries; (4) working out a sparse expression coefficient of an input low-resolution image block; (5) reestablishing a high-resolution image block by using a high-resolution dictionary and the spare coefficient; (6) performing weighting and summing on the high-resolution image block to obtain the high-resoluiton image block subjected to weighting and summing; (7) obtaining a high-resolution image according to the high-resolution image block; and (8) carrying out high-frequency information enhancement on the high-resolution image through error compensation to obtain a final result. A simulation experiment shows that the image super-resolution reconstruction method has the advantages of clear image surface and sharpened margin and can be used for image identification and target classification.

The invention discloses an integration method for face recognition by using sparse representation, which mainly solves the problem of low recognition stability in the conventional K-SVD dictionary learning method. The integration method is realized by the following steps of: generating a rotation matrix by a rotation forest algorithm; randomly projecting the same face sample data to different coordinate systems through the rotation matrix, wherein the projected face sample data is easier to distinguish than the original data; recognizing the projected face sample data by a sparse representation classification method; and voting to select the recognition result of a projected face sample to acquire the recognition result of the original face sample. Compared with the conventional sparse representation-based classification method, the integration method has the characteristics of improving the recognition correctness and the recognition stability, and can be used for a safety verification system.

The invention discloses a self-adaption online dictionary learning super-resolution method. The method firstly selects one group of high-resolution training image set and one group of low-resolution training image set,then establishes a relationship on the high-resolution training image set and the low-resolution training image set, and provides more prior information for image reconstructing through high-resolution/low-resolution dictionary to obtain a better reconstructing effect under the condition of high amplification factor, at the same time, the method uses SP (Self Propelled) sparse coding algorithm, overcomes the defect of low reconstructing accuracy due to traditional greedy algorithm and keeps a lower computation complexity. The method has the advantages of overcoming the defects of complex computation and low training speed in the current mainstream dictionary learning algorithms such as MOD (Multimedia On Demand) and K-SVD(k-singular value decomposition) and shortening image reconstructing time in general. Compared with the prior art based on learning image super-resolution, the method of the invention has a higher reconstructing accuracy and a shorter algorithm time.

The invention discloses a three-dimensional image quality objective evaluation method based on sparse representation. According to the method, in a training stage, left viewpoint images of a plurality of original undistorted three-dimensional images are selected for forming a training image set, Gaussian difference filtering is adopted for carrying out filtering on each image in the training image set to obtain filtered images in different scales, and in addition, a K-SVD method is adopted for carrying out dictionary training operation on a set formed by all sub blocks in all of the filtered images in different scales for constructing a visual dictionary table; and in a test stage, the Gaussian difference filtering is performed on any one tested three-dimensional image and the original undistorted three-dimensional image to obtain filtered images in different scales, then, the filtered images in different scales is subjected to non-overlapped partition processing, and an image quality objective evaluation prediction value of the tested images is obtained. The three-dimensional image quality objective evaluation method has the advantages that a complicated machine learning training process is not needed in the training stage; and the in the test stage, the image quality objective evaluation prediction value only needs to be calculated through a sparse coefficient matrix, and in addition, the consistency with the subjective evaluation value is better.

The invention relates to an image denoising method based on sparse regularization. The image denoising method based on sparse regularization fully utilizes the gradient information of an image and non local self-similarity to construct a sparse regularization denoising model, and utilizes an iteration histogram standardization algorithm to solve the model. The solving process includes: dividing the image into image blocks, and then clustering the image blocks according to the level of structural similarity, and at last using a sparse K-SVD dictionary to train each given image block. Therefore, the structural property of the dictionary is improved and also the over-complete dictionary obtained through training can preferably perform sparse representation of the image blocks. The image denoising method based on sparse regularization has the advantages of effectively denoising the image, being high in the reservation capability for the image texture structure, obtaining a better image visual effect, further reducing the computation complexity, and improving the operation speed.

The invention relates to a FiberBragg grating(FBG) signal self-adapting restoration method based on compressed sensing, and belongs to a signal restoration technology field of an optical fiber sensing system. The FBG signal self-adapting restoration method comprises steps that step 1: EMD combination mutual information is used for self-adapting denoising processing of spectral signals; step 2, segmented testing of a denoising signal is carried out, and the signal is divided into k segments, and sample databases corresponding to the signals are acquired by calculating Euclidean distances among various segments of signals and samples, and self-adapting dictionaries D corresponding to the signals are acquired by adopting a K-SVD dictionary learning method; step 3, measured signals are used to acquire observation matrixes R and observation signals xi; step 4, the observation signals are reconstructed by adopting an improved regularized orthogonal matching pursuit algorithm to acquire complete reconstructed signals. The FBG signal self-adapting restoration method is advantageous in that problems such as interferences of noises on the signals, targeted dictionary learning, and the signal self-adapting reconstruction are considered, and each part represents the self-adaptability of the algorithm, and can be flexibly used in practical engineering, and then influences caused by manual misoperation are reduced.

Disclosed is a single image super-resolution method based on identical scale structure self-similarity and compressed sensing. Firstly, the interpolation is performed for a low-resolution image and a quasi-high-resolution image is obtained; then, the quasi-high-resolution image is divided into quasi-high-resolution image blocks, vectors corresponding to the quasi-high-resolution image blocks serve as a training sample, a sample matrix is assembled, a K-SVD dictionary studying method is used for a solution and a dictionary is obtained; the low-resolution image is divided into low-resolution image blocks; by the aid of a down-sampling matrix, the dictionary and vectors corresponding to all low-resolution image blocks, an orthogonal matching pursuit (OMP) method is used for a solution, and vectors corresponding to high-resolution reconstruction image blocks; and finally, vectors corresponding to high-resolution reconstruction image blocks are assembled and a high-resolution reconstruction image is formed. According to the super-resolution method based on the identical scale structure self-similarity and the compressed sensing, additional information is added in the high-resolution reconstruction image through a compressed sensing frame, and the space resolution is improved.

The invention discloses an infrared image preprocessing method based on the compressed sensing algorithm. The method comprises the steps of conducting primary processing on original infrared image information to obtain processed image information, generating a redundant dictionary according to the processed image information, conducting sparse representation on the processed image under the redundant dictionary by means of the orthogonal matching algorithm to obtain a sparse matrix of the processed image, training the redundant dictionary and the sparse matrix by means of the K-SVD algorithm to obtain a trained redundant dictionary and a trained sparse matrix, and obtaining a final infrared image signal through calculation according to the trained redundant dictionary and the trained sparse matrix. According to the method, the infrared image is preprocessed by means of the compressed sensing algorithm, and the signal to noise ratio of the processed infrared image is increased greatly.

The invention provides a point cloud data sparse representation method based on compressed sensing. Under the premise of ensuring certain precision, massive point cloud data is compressed, thus the sparsity of the point cloud data is greatly raised, and a good foundation is laid for point cloud data compression and reconstruction based on compressed sensing. The method comprises the steps of (1) point cloud data normalization, (2) over-complete dictionary sparse representation based on a K-SVD algorithm, (3) normalized point cloud data observation, transmission and storage, (4) point cloud data reconstruction based on l1 norm minimization, and (5) normalized point cloud data recovery.

A single-image super-resolution method based on the multi-scale structural self-similarity and the compressive sensing comprises the following steps of: firstly setting an initial estimated value of a high resolution reconstructed image, setting a stopping error and the maximum time of iteration, determining a downsampling matrix and a fuzzy matrix according to the process of image degradation to construct an image pyramid, and building a dictionary by using the image pyramid as a training sample of the K-SVD (K-singular value decomposition) method; secondly, according to a Nonlocal method, searching for similar image blocks with the same scale in the current high resolution reconstructed image and determining a weight matrix; thirdly, updating the estimated value of the high resolution reconstructed matrix, updating the sparse representation coefficient, and updating the estimated value of the high resolution reconstructed matrix again; and fourthly carrying out the next iteration until two sequential high resolution reconstructed matrixes meet the corresponding requirement or reach the maximum time of iteration. The single-image super-resolution method of the invention adds the additional information contained in a multi-scale self-similar structure of an image into the high resolution reconstructed image through a compressive sensing frame, thereby having a high computational efficiency.

The invention discloses a K-SVD (K-means singular value decomposition) speckle inhibiting method based on SAR (synthetic aperture radar) image local statistic characteristic, mainly solving the problem that detail information such as edge, texture and the like is fuzzy in the traditional speckle inhibiting method. The method is realized in the following processes of: inputting an SAR image, extracting overlapped blocks in the SAR image to obtain an overlapped block vector set; then randomly sampling the overlapped block vector set to obtain a training sample set; carrying out SAR_KSVD dictionary training on a training sample to obtain a final training dictionary; carrying out SAR_OMP sparse coding on the overlapped block vector set under the condition of the final training dictionary to obtain a sparse coding coefficient; and obtaining a speckle inhibited image by utilizing the final training dictionary and the sparse coding coefficient according to the redundant sparse representation image noise inhibiting theory. By applying the method disclosed by the invention, speckle noise in a homogenous region can be effectively inhibited, brightness and edge texture of a target at a strong reflection point can be well maintained to be clear, and the method disclosed by the invention can be applicable to SAR images in the fields such as land resource monitoring, natural disaster analysis and the like.

The invention discloses a shared k-SVD dictionary-based DVS visual video denoising method. The problem that an image object generated at a high frame rate is unclear in contour and long in consumed time is mainly solved. The scheme of the method is as follows: 1, installing the drive of a dynamic video sensor, capturing an event stream and storing the event stream; 2, converting the event stream into DVS images with clear contours, and grouping the images; 3, obtaining an optimized dictionary of a first frame of image in each group according to the k-SVD algorithm, and subjecting all the restimages to denoising treatment by using a learning dictionary obtained by the first frame of image in each group; 4, setting a video frame rate and a frame number, and carrying out transcoding treatment on the DVS images after being subjected to denoising treatment. According to the invention, an object has obvious profile and other characteristics under the condition that a high frame rate is ensured. Moreover, a good denoising effect and a high denoising speed can be achieved while the structure information of the object is reserved. The method can be used for the image pretreatment of DVS development.

The invention discloses a depth image denoising method which can reduce edge blur of images, is applicable to non-zero-mean Gaussian denoising and greatly improves depth image quality. The depth image denoising method includes the steps that 1, joint bilateral filtering is carried out on the depth images, the action scope of a bilateral filter is constrained, and accordingly filtered images are obtained; 2, a dictionary is trained through a K-SVD method, the filtered images of the first step are denoised through sparse representation on the basis of the dictionary, and reconstructed images are obtained.

The invention discloses a deep learning-based quantitative modeling method of near-infrared spectroscopy of tobacco. A near-infrared spectrometer is utilized to collect spectroscopy information, the near-infrared spectroscopy information of the tobacco is acquired, and spectroscopy data are preprocessed; main chemical component information of the tobacco is acquired; a sparse feature learning method is used to create an over-complete dictionary by applying the near-infrared spectroscopy data of the tobacco and a K-SVD algorithm, and an OMP algorithm is utilized to calculate and obtain sparse representation coefficients of spectroscopy; and a PSO-SVM learning algorithm is adopted, and the sparse representation coefficients and the chemical component information are combined to establish a near-infrared spectroscopy regression prediction model. According to the method, dual technologies of spectroscopy analysis and machine learning are utilized, and a support vector machine algorithm in pattern recognition is combined to realize fast quantitative modeling for the near-infrared spectroscopy of the tobacco, and the established model is applied to accurately predicting the main chemical component information of the tobacco.

The invention relates to a discriminative dictionary learning based multi-source image fusion denoising method. The method includes: acquiring multi-source images as training samples, and learning thesamples through a K-SVD algorithm to obtain an initial cartoon dictionary and an initial texture dictionary, introducing weighting nuclear norm constraint to bring forward a new dictionary learning model, performing new dictionary learning model learning to obtain a cartoon dictionary and a texture dictionary, decomposing to-be-fused images through an MCA algorithm to obtain a cartoon component and a texture component, introducing weighting Schatten sparse nuclear norm constraint to the cartoon component, adding grey level histogram gradient protection to the texture component to bring forward a new image decomposition model, iterating the model to obtain a cartoon sparse coding coefficient and a texture sparse coding coefficient, respectively fusing to obtain a cartoon component and a texture component according to a principle of maximum sparse coding coefficient l1 norm values of corresponding components, and adding the cartoon component and the texture component to obtain a final fusion image. The method has advantages that image fusion and denoising are realized, false information transferring is avoided, time consumption is reduced, and fusion and denoising performances are improved.

The invention discloses a K-SVD and sparse representation based polarization SAR (synthetic aperture radar) image classification method and solves problems that the amount of classification categories is limited by an existing method and polarization characteristic information is not fully utilized. The method includes the steps of 1), calculating a covariance matrix by taking a polarization coherent matrix of a polarization SAR as input data; 2), extracting the coherent matrix, the covariance matrix, Ps, Pd, Pv, H, and alpha from each pixel to form a characteristic matrix; 3), selecting training samples from actual terrain distribution to form an initial dictionary; 4), training the initial dictionary with a K-SVD algorithm to obtain a training dictionary; 5), representing the characteristic matrix with the training dictionary and calculating the sparse coefficient with an OMP algorithm; 6), restructuring the characteristic matrix with the calculated sparse coefficient, and determining the categories of pixels to acquire the final classification result. Polarization characteristics of polarization SAR images are utilized, the amount of the classification categories is not limited, and the method can be applied to classification of the polarization SAR images.

The invention discloses an image noise reduction system and method based on K-SVD (Singular Value Decomposition) and locally linear embedding, and particularly relates to a signal sparse representation and reconstruction technique based on dictionary learning and an image noise reduction technique based on manifoid learning. A K-SVD method is adopted to be a frame as the signal sparse representation and reconstruction technique based on the dictionary learning, the locally linear embedding used as a constrain condition is added into a target function while the signal sparse representation is solved, so that the relation of the decomposed sparse coefficients is strengthened, the influence of the random noise on the sparse coefficients is overcome, and the image noise reduction effect better than that of the original K-SVD method is obtained.