Sparse Representation-based Classification Method Research Articles

Clustering-weighted SIFT-based classification method via sparse representation

In recent years, sparse representation-based classification (SRC) has received significant attention due to its high recognition rate. However, the original SRC method requires a rigid alignment, which is crucial for its application. Therefore, features such as SIFT descriptors are introduced into the SRC method, resulting in an alignment-free method. However, a feature-based dictionary always contains considerable useful information for recognition. We explore the relationship of the similarity of the SIFT descriptors to multitask recognition and propose a clustering-weighted SIFT-based SRC method (CWS-SRC). The proposed approach is considerably more suitable for multitask recognition with sufficient samples. Using two public face databases (AR and Yale face) and a self-built car-model database, the performance of the proposed method is evaluated and compared to that of the SRC, SIFT matching, and MKD-SRC methods. Experimental results indicate that the proposed method exhibits better performance in the alignment-free scenario with sufficient samples.

Face recognition using position-dictionaries and region covariance feature

In this paper, a new sparsity formulation called position-dictionary based sparse representation is developed for frontal face recognition. Different from the sparse representation based classification (SRC) method and the Gabor-feature based SRC (GSRC) method which both employ a global dictionary to decompose image patches, the proposed method constructs a position-dictionary for each location using training patches in the corresponding location since they resemble each other and are more likely to favor the same atoms. Sparse coefficients of each position-patch can be obtained by solving an \(l_{1}\)-norm minimization problem. For each face image, sparse coefficients of position-patches are pooled to construct a discriminative upper level feature to represent face image. PCA is used to perform dimension reduction. Each testing sample is represented as a sparse linear combination of all training samples, and recognition is accomplished by evaluating which class of training samples leads to the minimum reconstruction error. We compared the proposed method with SRC and GSRC method on three benchmark face databases. Experimental results show that the proposed method achieves higher recognition rates and is robust to a certain degree of occlusions.

Weighted group sparse representation for undersampled face recognition

Abstract Recently, sparse representation has become a popular data representation method and sparse representation based classification (SRC) has been proposed. However, the l1-norm regularized sparse representation is not stable and fails to incorporate the label information of training samples. On the other hand, the training samples are grouped into a column-based matrix, which has a group structure. In this paper, we propose a new classification method called weighted group sparse representation classification (WGSRC) to classify a query image by minimizing the weighted mixed-norm (l2,1-norm) regularized reconstruction error with respect to training images. Unlike SRC, the group sparse regularization is utilized to incorporate the label information and it promotes sparsity at the group level. According to the similarity between a test sample and training samples of each group, WGSRC gives each group a weight. Our method integrates the locality structure of the data and similarity information between the query sample and distinct classes into l2,1-norm regularization to form a unified formulation. We try to represent a test sample by training samples not only from the neighbors of it, but also from the highly relevant classes. The sparse solution of WGSRC encodes more structure information and discriminative information than other sparse representation methods. Experimental results on five face data sets have shown that the proposed method outperforms state-of-the-art sparse representation based classification methods, especially when we have relatively small number of training data for each class. We also explore the reasons why the proposed approach can improve face recognition.

EPILEPTIC EEG CLASSIFICATION BASED ON KERNEL SPARSE REPRESENTATION

The automatic identification of epileptic EEG signals is significant in both relieving heavy workload of visual inspection of EEG recordings and treatment of epilepsy. This paper presents a novel method based on the theory of sparse representation to identify epileptic EEGs. At first, the raw EEG epochs are preprocessed via Gaussian low pass filtering and differential operation. Then, in the scheme of sparse representation based classification (SRC), a test EEG sample is sparsely represented on the training set by solving l1-minimization problem, and the represented residuals associated with ictal and interictal training samples are computed. The test EEG sample is categorized as the class that yields the minimum represented residual. So unlike the conventional EEG classification methods, the choice and calculation of EEG features are avoided in the proposed framework. Moreover, the kernel trick is employed to generate a kernel version of the SRC method for improving the separability between ictal and interictal classes. The satisfactory recognition accuracy of 98.63% for ictal and interictal EEG classification and for ictal and normal EEG classification has been achieved by the kernel SRC. In addition, the fast speed makes the kernel SRC suit for the real-time seizure monitoring application in the near future.

Recent progress in face recognition based on sparse coding

Sparse coding has been attracting increasing interests in computer vision filed, due to its adaptive learning ability and biological inspiration from human vision system. Since sparse representation based classification method for face recognition got great success in 2009, many subsequent improved methods were proposed. In this paper, we aim at providing a comprehensive review of the recent state-of-the-art face recognition methods based on sparse coding. By analyzing their advantages and disadvantages, we summarize the roles of sparse coding in face recognition and discuss the potential improvements in the future.

Local Spatiotemporal Coding and Sparse Representation Based Human Action Recognition

To handle with the limitation of bag-of-features (BoF) model which ignores spatial and temporal relationships of local features in human action recognition in video, a Local Spatiotemporal Coding (LSC) is proposed. Rather than the exiting methods only uses the feature appearance information for coding, LSC encodes feature appearance and spatiotemporal positions information simultaneously with vector quantization (VQ). It can directly models the spatiotemporal relationships of local features in space time volume (STV). In implement, the local features are projected into sub-space-time-volume (sub-STV), and encoded with LSC. In addition a multi-level LSC is also provided. Then a group of sub-STV descriptors obtained from videos with multi-level LSC and Avg-pooling are used for action video classification. A sparse representation based classification method is adopted to classify action videos upon these sub-STV descriptors. The experimental results on KTH, Weizmann, and UCF sports datasets show that our method achieves better performance than the previous local spatiotemporal features based human action recognition methods.

Scale invariant texture classification via sparse representation

Scale change exists very commonly in real-world textural images which remains one of the biggest challenges in texture classification due to the tremendous changes involved in texture appearance. While most research efforts have been devoted to extracting various scale invariant features, these methods are either unsuitable to describe a texture or unable to handle the situations where a large amount of scale change exists. Other works attempt to avoid scale invariant feature extraction by generating a set of multi-scale representations from training images for classification, but they are not only computation intensive but also limited to dealing with small scale changes between training images and test images. In this paper we investigate the scaling properties of textures and introduce a low dimensional linear subspace for the multi-scale representations of a texture, in which the collaboration between the multi-scale representations is beneficial for the scale invariant texture classification. We therefore propose a new scale invariant texture classification framework without extracting scale invariant features, by using a sparse representation technique to model the multi-scale representations of a texture and taking the advantages of collaboration between them for classification. Specifically, a multi-scale dictionary is constructed from the Gaussian-pyramid-generated scale space of a small set of training images at one scale, and then the test images at arbitrary scales are classified via a modified sparse representation based classification method. Experiments on two benchmark texture databases show that the proposed method is able to deal with large scale changes between the training images and the test images and achieve comparative results to the state-of-the-art approaches for the classification of textures with various variations, especially scale.

A new sparse representation-based classification algorithm using iterative class elimination

In this paper, we propose a sparse representation-based classification method using iterative class elimination strategy for face recognition. The proposed method aims to represent a test sample as a linear combination of the most competitive training samples and exploits an optimal representation of training samples from the classes with major relevant contributions. We interpret the sparse representation problem as an information fidelity problem. In the context of our proposed method, an important goal is to select a subset of variables for accomplishing one objective: the provision of a descriptive representation for sparse class knowledge structure. We develop an iterative class elimination algorithm to achieve this goal. First, the contribution in presenting the test sample of any of the specified classes is, respectively, calculated by adding up the total contribution of all training samples of this class, and then a certain class that meets the smallest score requirement to this test sample is eliminated. Second, a similar procedure is iteratively carried out for the set of remaining training samples from rest classes, and this procedure is repeatedly performed till the predefined termination condition is satisfied. The final remaining training samples are used to produce a best representation of the test sample and to classify it. Therefore, the proposed algorithm is an iterative method that alternates between sparse representation and a process of updating the training atoms to better fit the test data. This is helpful to accurately classify the test sample. Experimental results conducted on the ORL, FERET, and AR face databases demonstrate the effectiveness of the proposed method.

Locality-sensitive dictionary learning for sparse representation based classification

Motivated by image reconstruction, sparse representation based classification (SRC) has been shown to be an effective method for applications like face recognition. In this paper, we propose a locality-sensitive dictionary learning algorithm for SRC, in which the designed dictionary is able to preserve local data structure, resulting in improved image classification. During the dictionary update and sparse coding stages in the proposed algorithm, we provide closed-form solutions and enforce the data locality constraint throughout the learning process. In contrast to previous dictionary learning approaches utilizing sparse representation techniques, which did not (or only partially) take data locality into consideration, our algorithm is able to produce a more representative dictionary and thus achieves better performance. We conduct experiments on databases designed for face and handwritten digit recognition. For such reconstruction-based classification problems, we will confirm that our proposed method results in better or comparable performance as state-of-the-art SRC methods do, while less training time for dictionary learning can be achieved.

Collaborative neighbor representation based classification using [formula omitted]-minimization approach

Applications of sparse signal representation in image processing and pattern recognition have attracted a great deal of attention. Sparse representation based classification (SRC) methods emphasizes on sparse representation computed by l1-minimization to exploit the underlying sparsity in the problem domain, and argued the importance of sparse representation that improved the discrimination to achieve robust and accurate classification results. Recently, many studies have shown the role of collaborative representation (CR) in SRC, which actually improved the classification accuracy. In this paper, we proposed a novel collaborative neighbor representation method for multi-class classification based on l2-minimization approach with the assumption of locally linear embedding (LLE). The proposed method represents a test sample over the dictionary by automatically choosing optimal nearest basis spanned in the same linear subspace as of test sample. The proposed representation method achieves competitive classification accuracy via optimal neighbor representation having discriminative learning power. Extensive experiments on real-world face and digit databases are performed to analyze the performance of the proposed method against SRC methods. Result clearly shows that the proposed method achieves competitive results for face recognition and pattern classification, and is significantly much faster and comparably accurate than SRC based classification methods.

Enhancing sparsity via ℓp (0

Sparse representation has received an increasing amount of interest in recent years. By representing the testing image as a sparse linear combination of the training samples, sparse representation based classification (SRC) has been successfully applied in face recognition. In SRC, the ℓ1 minimization instead of the ℓ0 minimization is used to seek for the sparse solution for its computational convenience and efficiency. However, ℓ1 minimization does not always yield sufficiently sparse solution in many practical applications. In this paper, we propose a novel SRC method, namely the ℓp (0<p<1) sparse representation based classification (ℓp-SRC), to seek for the optimal sparse representation of a testing image. In ℓp-SRC, the ℓp (0<p<1) minimization is adopted as an alternative to ℓ0 minimization, the solution of which is sparser than that of ℓ1 minimization used in traditional SRC. Furthermore, an iterative algorithm is introduced to efficiently solve the ℓp minimization problem in this paper. The extensive experimental results on publicly available face databases demonstrate the effectiveness of ℓp-SRC for robust face recognition.

Sparse representation-based classification scheme for motor imagery-based brain–computer interface systems

Motor imagery (MI)-based brain–computer interface systems (BCIs) normally use a powerful spatial filtering and classification method to maximize their performance. The common spatial pattern (CSP) algorithm is a widely used spatial filtering method for MI-based BCIs. In this work, we propose a new sparse representation-based classification (SRC) scheme for MI-based BCI applications. Sensorimotor rhythms are extracted from electroencephalograms and used for classification. The proposed SRC method utilizes the frequency band power and CSP algorithm to extract features for classification. We analyzed the performance of the new method using experimental datasets. The results showed that the SRC scheme provides highly accurate classification results, which were better than those obtained using the well-known linear discriminant analysis classification method. The enhancement of the proposed method in terms of the classification accuracy was verified using cross-validation and a statistical paired t-test (p < 0.001).

Prostate segmentation by sparse representation based classification.

Accurate segmentation of prostate in CT images is important in image-guided radiotherapy. However, it is difficult to localize the prostate in CT images due to low image contrast, unpredicted motion and large appearance variations across different treatment days. To address these issues, we propose a sparse representation based classification method to accurately segment the prostate. The main contributions of this paper are: (1) A discriminant dictionary learning technique is proposed to overcome the limitation of the traditional Sparse Representation based classifier (SRC). (2) Context features are incorporated into SRC to refine the prostate boundary in an iterative scheme. (3) A residue-based linear regression model is trained to increase the classification performance of SRC and extend it from hard classification to soft classification. To segment the prostate, the new treatment image is first rigidly aligned to the planning image space based on the pelvic bones. Then two sets of location-adaptive SRCs along two coordinate directions are applied on the aligned treatment image to produce a probability map, based on which all previously segmented images of the same patient are rigidly aligned onto the new treatment image and majority voting strategy is further adopted to finally segment the prostate in the new treatment image. The proposed method has been evaluated on a CT dataset consisting of 15 patients and 230 CT images. Promising results have been achieved.