Sparse Representation Classification Research Articles

HM–GDM: Hybrid Measures and Graph-Dependent Modeling for Environmental Sound Classification

One of the upcoming areas of research is the automated environmental sound classification, as its application is highly utilized in criminal investigations scenarios, surveillance systems, biomedical fields, radio navigation systems, etc. The procedure of environmental sound classification deals with standardized techniques such as feature extraction and feature selection, followed by classification with machine learning or deep learning techniques. In this work, five feature extraction techniques are used initially such as the discrete wavelet transform (DWT)-based empirical mode decomposition (EMD) technique, phase space reconstruction (PSR), kernel principal component analysis (KPCA), variational mode decomposition (VMD) and tunable Q-wavelet transform (TQWT). The extracted features are then selected through seven feature selection techniques, out of which two have been proposed newly and five have already existed. The proposed feature selection techniques are stratified clustered collective technique (SCCT) and fuzzy-based template shape clustering (FTSC). The other feature selection techniques used are the generalized discriminant analysis (GDA), ReliefF, Fisher discriminant criterion (FDC), Kruskal–Wallis test and coati optimization algorithm (COA). The selected features are then classified through the proposed swarm intelligence-based hybrid Adaboost – random forest termed as SIHAR classifier and the selected features are classified with the less explored sparse representation classifier (SRC) along with eight other traditionally used machine learning classifiers. The work also proposes graph-dependent modeling for the environmental sounds where the rhythms are extracted for the similarity assessment and later the decision-making is done using hypothesis testing. The work is tested on Firat ESC-50 dataset and the best results are produced in terms of a high classification accuracy of 87.48% which is obtained for the TQWT + SCCT + SIHAR combination.

Platform Design of Passive Target Perception and Localization Based on Sensor Networks

With the rapid development, extensive knowledge, and diverse application scenarios of target perception and positioning technology in sensor networks, a passive target perception and localization platform based on Wireless Sensor Networks (WSN) has been designed. The platform is mainly applied for the teaching of electronic information, communication, and Internet of Things (IoT) engineering. The platform follows the teaching concept of “developing students’ ability to solve complex engineering problems” in the emerging engineering discipline and combines experimental simulation with real-world testing, as well as integrating scientific research and teaching. It encompasses technical elements such as perception and localization theory, sparse representation classification modeling, solving underdetermined equations, data analysis, and sparse coding, as well as nontechnical aspects such as team collaboration and cost budgeting. The platform boasts high fidelity and scalability, providing students with the opportunity to develop comprehensive practical and innovative skills in solving complex engineering problems.

The multi-channel signals based tensor sparse representation classification method for fault diagnosis of high-speed train

The transmission system is a key component to ensure the stable operation of high-speed trains. Thus, it is significant to monitor its condition to ensure the operation safety. Nowadays sparse representation is widely used in fault diagnosis. However, as the number of sensors is increasing, the existing method destroys the internal structure of multi-channel signals and cannot effectively deal with the fault diagnosis of multi-channel signals in parallel. Therefore, this article extends the existing sparse representation method to tensor space to extract the coupling information between channels and realize the fault diagnosis of multi-channel. First, a tensor sparse representation model is proposed to achieve data-level multi-channel signal fusion and complete inter-channel fault feature extraction. Then, a multimodal dictionary learning algorithm is proposed to adaptively design the data-driven dictionary to achieve data-driven feature extraction. Finally, a tensor sparse representation classification method is proposed to achieve the purpose of intelligent diagnosis. Fault experiments verify the effectiveness and superiority of the method.

A weighted sparse classification method based on period analysis dictionary

The periodic transient shocks triggered by damages in rolling bearings are frequently overshadowed by disruptive elements such as noise and harmonics. Therefore, the extraction of fault characteristics from these disturbances to identify the health status of the bearing is crucial for fault diagnosis. This study presents a novel approach, the period analysis dictionary weighted sparse representation classification (PAD-WSRC) method, designed specifically for rolling bearings. The proposed approach incorporates Bi-damped wavelet as the dictionary wavelet atom, while accounting for the pulse characteristics induced by faults and leveraging prior knowledge of periodicity. A weighted sparse representation (SR) method was also designed, which calculates the weighted SR coefficients of samples, amplifying the local features of samples while addressing the impact of time-shift bias. In addition, the bearing fault type is identified using a mutual correlation classification criterion based on sparse approximation. Our PAD-WSRC strategy has demonstrated its effectiveness in classifying the health status of bearings across three datasets, achieving recognition accuracies of 99.75%, 99.69% and 99.17%, respectively. Comparisons with several traditional methods further underscore the effectiveness and superiority of our proposed method in classifying rolling bearing faults.

Research on Mushroom Image Classification Algorithm Based on Deep Sparse Dictionary Learning

The traditional mushroom feature extraction method has low classification efficiency and unsatisfactory effect. Dictionary learning is widely used in image classification. However, the previous work is to learn dictionaries in the original space, which limits the performance of sparse representation classification. In order to solve the problem of spatial redundancy in traditional convolutional neural networks and the weak performance of deep learning in small samples, an improved dictionary learning algorithm, Deep Sparse Dictionary learning (DSDL), is proposed. The input to DSDL is not a matrix gathered from the original grayscale image or a hand-created feature, but rather a relatively deeper feature extraction via a stack autoencoder. Then, a structured dictionary is designed to reconstruct the deep features according to different categories of distinguishing features. In addition, it is necessary to learn the associated structured projection sparse dictionary to ensure that the decoder updates in the direction of the deconvolution operator error is minimal. By utilizing sparse dictionary learning loss functions and autoencoder loss functions, DSDL can simultaneously learn deep latent features and corresponding dictionary pairs. In the testing phase of DSDL, the minimum errors of deep feature and structured projection components for different classes can be directly represented by basic matrix multiplication operations. Experimental results show that the proposed method achieves a good classification effect on mushroom images, which shows the effectiveness of the method.

A novel adaptive generalized domain data fusion-driven kernel sparse representation classification method for intelligent bearing fault diagnosis

Dictionary learning has gradually attracted attention due to its powerful feature representation ability. However, the time-shift property of collected signals hinders the recognition of various bearing states. In addition, existing dictionary learning methods are mostly designed based on a single domain, while common data fusion methods used in data-driven cannot be directly extended to dictionary learning. In this paper, a novel adaptive generalized domain data fusion-driven kernel sparse representation classification method (AGDFDK-SRC) is proposed. First, to avoid the effect of the time-shift property on dictionary learning, a class-specific kernel sub-dictionary learning method is proposed, by which the non-linear signal data is mapped into high-dimensional feature space via a kernel trick. Second, the class-specific kernel sub-dictionaries are learned by kernel K-singular value decomposition in a data-driven manner. Then, an adaptive generalized domain data fusion strategy is developed for dictionary learning, which implements data fusion of multiple domain signals to enhance the feature mining ability and representation ability of the learned dictionary. Finally, a kernel sparse classification method is designed to achieve intelligent bearing fault diagnosis. Two bearing datasets are exploited to verify the recognition performance of AGDFDK-SRC, indicating that the AGDFDK-SRC obtains superior average classification accuracies of 98.23% and 99.50%, respectively.

EEG-based schizophrenia classification using penalized sequential dictionary learning in the context of mobile healthcare

Mobile healthcare is an appealing approach based on the Internet of Medical Things (IoMT) and cloud computing. It can lead to unobstructed, economical, and patient-centric healthcare solutions. The key performance indicators of such systems are dimensionality reduction, computational effectiveness, low latency, and accuracy. In this context, a novel approach is devised for EEG-based schizophrenia, a severe mental disorder that adversely affects a person’s behavior and classification. A multichannel EEG recording with suitable granularity is required for precise analysis. It can increase exponentially the data dimensionality plus complexity and computational load. The proposed solution attains an interesting trade-off between dimensionality reduction plus computational effectiveness versus accuracy. It uses the penalized sequential dictionary learning (PSDL) that incorporates channel selection. First, PSDL learns a dictionary from the input data and evaluates its performance on all EEG channels. Based on this evaluation, a subset of six channels is selected for further training in the dictionary. The proposed PSDL algorithm then incorporates a penalty term that enhances the power of the learned dictionary on the selected channels. We evaluate the proposed approach on the multi-channel EEG dataset from the Institute of Psychiatry and Neurology in Warsaw, Poland. A performance comparison is also made with counterparts. The models’ performance depends on the EEG signals’ complexity. Therefore, we tried to make our models robust and straightforward, achieving appropriate performance with minimal computational cost. The proposed method reduces the dimension in two steps. First, the count of channels is reduced to 68.42%. In the second step, the kept information, 31.58% of channels, is further reduced to 83.75% using dictionary learning. The proposed framework secures a remarkable data dimension reduction and a lower computational cost and latency than the counterparts while attaining the sparse representation classification accuracy of 89.12%. These findings are promising and confirm the potential of investing in incorporating the proposed method in contemporary mobile healthcare solutions.

Embedding vector field into image features to enhance classification

Classical methods, including sparse representation classification (SRC) and neural networks (NNs), classify image object(s) using features like intensity, color, texture, and geometry extracted from the image. For the purpose of classification the present study proposes to augment the set of image features with vector field (VF) features, like singular points (SPs) and trajectories. To generate VFs with such features, our approach solves the Poisson Image (PI) equation. Using its solution uˆ(x,y), our approach defines two functions ϕˆ(x,y) and ψˆ(x,y) and develops two Poisson gradient VFs (PGVFs) ∇ϕˆ(x,y) and ∇ψˆ(x,y), used to embed features into a database of images. The embedding of VF features into image objects constitutes the main novelty of this study. The advantage that comes from the novelty is increased statistics of machine learning (ML) classifiers. To validate the advantage, the PGVFs ∇ϕˆ(x,y) and ∇ψˆ(x,y) were embedded into the public image databases COIL100, YaleB, ISIC2018, and ISIC2020. The original and the eight VF covered databases were classified with two ML classifiers: sparse representation wavelet classification (SRWC) and SRC in the quaternion wavelets (SRCQW) domain. The results obtained are presented in the paper and confirm the claimed advantage.

Effective superpixel sparse representation classification method with multiple features and L0 smoothing for hyperspectral images

Effective superpixel sparse representation classification method with multiple features and L0 smoothing for hyperspectral images

A hybrid SVM and kernel function-based sparse representation classification for automated epilepsy detection in EEG signals

Automatic epilepsy detection based on electroencephalography (EEG) is crucial for advancing the diagnosis and treatment of epilepsy. In this paper, we propose a novel classification algorithm called SVM-KSRC, which differs from integrated learning approaches. The algorithm establishes a connection between support vector machine (SVM) and kernel sparse representation classification (KSRC) using support vectors. Specifically, we extract two types of features from the pre-processed EEG signals in this study. During the training phase, these features are utilized to train the SVM model and construct the kernel sparse representation dictionary. We differentiate the SVM part of the features of the test data to determine whether SVM or KSRC should be employed for classifying the test data. Our method is evaluated on two publicly available datasets: University of Bonn dataset and Neurology and Sleep Centre-New Delhi dataset. Through 10 times 10-fold cross validation, our method demonstrates superior performance in epilepsy detection when compared to existing machine learning methods. The experimental results demonstrate that SVM-KSRC is more effective compared to SVM and KSRC used separately. It achieves over 99% accuracy in all binary classification tasks and attains 100% accuracy in certain tasks. The source code is publicly available at https://github.com/Walkeraaa/SVM-KSRC.

Spectral ensemble sparse representation classification approach for super-robust health diagnostics of wind turbine planetary gearbox

Health monitoring, diagnostics and prognostics techniques have been deemed as the most promising and essential framework towards smart operation and maintenance of wind energy equipment. Wind turbine planetary gearboxes have remained the most intricate and challenging transmission units to implement intelligent health diagnostics in wind power generation systems. To resolve this issue, we present a novel spectral ensemble sparse representation classification (S-ESRC) approach for super-robust health diagnostics of wind turbine planetary gearboxes. Specifically, S-ESRC implements super-robust health diagnostics via three procedures consisting of data augmentation, spectral dictionary design, and spectral sparse approximation-based diagnostic scheme. Firstly, the prediction translation-invariance is exploited to accomplish vibrational data augmentation. Second, the spectral dictionary design with robust and strong reconstruction capability is achieved via spectrum construction and feature fusion considering the intra-class and inter-class attributes. Thirdly, the spectral sparse approximation error-based diagnostic scheme is applied to accomplish robust health diagnostics. Experimental validations using a wind turbine planetary gearbox system have demonstrated the applicability and superiority of S-ESRC for super-robust health diagnostics. Comparative studies have comprehensively shown the super-robust performances of S-ESRC including superior diagnostic accuracy, strong robustness to random noises, strong robustness to hyperparameters, and efficient computation costs in comparison with several state-of-the-art approaches.

Motor imagery classification using sparse representations: an exploratory study

The non-stationary nature of the EEG signal poses challenges for the classification of motor imagery. sparse representation classification (SRC) appears as an alternative for classification of untrained conditions and, therefore, useful in motor imagery. Empirical mode decomposition (EMD) deals with signals of this nature and appears at the rear of the classification, supporting the generation of features. In this work we evaluate the combination of these methods in a multiclass classification problem, comparing them with a conventional method in order to determine if their performance is regular. For comparison with SRC we use multilayer perceptron (MLP). We also evaluated a hybrid approach for classification of sparse representations with MLP (RSMLP). For comparison with EMD we used filtering by frequency bands. Feature selection methods were used to select the most significant ones, specifically Random Forest and Particle Swarm Optimization. Finally, we used data augmentation to get a more voluminous base. Regarding the first dataset, we observed that the classifiers that use sparse representation have results equivalent to each other, but they outperform the conventional MLP model. SRC and SRMLP achieve an average accuracy of 75.95% and 82.51% respectively while the MLP is 72.38%, representing a gain between 4.93% and 14%. The use of EMD in relation to other feature processing techniques is not superior. However, EMD does not influence negatively, there is an opportunity for improvement. Finally, the use of data augmentation proved to be important to obtain relevant results. In the second dataset, we did not observe the same results. Models based on sparse representation (SRC, SRMLP, etc.) have on average a performance close to other conventional models, but without surpassing them. The best sparse models achieve an average accuracy of 95.43% among the subjects in the base, while other model reach 98.33%. The improvement of self-adaptive mechanisms that respond efficiently to the user’s context is a good way to achieve improvements in motor imagery applications. However, other scenarios should be investigated, since the advantage of these methods was not proven in all datasets studied. There is still room for improvement, such as optimizing the dictionary of sparse representation in the context of motor imagery. Investing efforts in synthetically increasing the training base has also proved important to reduce the costs of this group of applications.

HISET: Hybrid interpretable strategies with ensemble techniques for respiratory sound classification

The human respiratory systems can be affected by several diseases and it is associated with distinctive sounds. For advanced biomedical signal processing, one of the most complex issues is automated respiratory sound classification. In this research, five Hybrid Interpretable Strategies with Ensemble Techniques (HISET) which are quite interesting and robust are proposed for the purpose of respiratory sounds classification. The first approach is termed as an Ensemble GSSR technique which utilizes L2 Granger Analysis and the proposed Supportive Ensemble Empirical Mode Decomposition (SEEMD) technique and then Support Vector Machine based Recursive Feature Elimination (SVM-RFE) is used for feature selection and followed by classification with Machine Learning (ML) classifiers. The second approach proposed is the implementation of a novel Realm Revamping Sparse Representation Classification (RR-SRC) technique and third approach proposed is a Distance Metric dependent Variational Mode Decomposition (DM-VMD) with Extreme Learning Machine (ELM) classification process. The fourth approach proposed is with the usage of Harris Hawks Optimization (HHO) with a Scaling Factor based Pliable Differential Evolution (SFPDE) algorithm termed as HHO-SFPDE and it is classified with ML classifiers. The fifth or the final approach proposed analyzes the application of dimensionality reduction techniques with the proposed Gray Wolf Optimization based Support Vector Classification (GWO-SVC) and another parallel approach utilizes a similar kind of analysis with the Grasshopper Optimization Algorithm (GOA) based Sparse Autoencoder. The results are examined for ICBHI dataset and the best results are shown for the 2-class classification when the analysis is carried out with Manhattan distance-based VMD-ELM reporting an accuracy of 95.39%, and for 3-class classification Euclidean distance-based VMD-ELM reported an accuracy of 90.61% and for 4-class classification, Manhattan distance-based VMD-ELM reported an accuracy of 89.27%.

Single-Sample Finger Vein Recognition via Competitive and Progressive Sparse Representation

As an emerging biometric technology, finger vein recognition has attracted much attention in recent years. However, single-sample recognition is a practical and longstanding challenge in this field, referring to only one finger vein image per class in the training set. In single-sample finger vein recognition, the illumination variations under low contrast and the lack of information of intra-class variations severely affect the recognition performance. Despite of its high robustness against noise and illumination variations, sparse representation has rarely been explored for single-sample finger vein recognition. Therefore, in this paper, we focus on developing a new approach called Progressive Sparse Representation Classification (PSRC) to address the challenging issue of single-sample finger vein recognition. Firstly, as residual may become too large under the scenario of single-sample finger vein recognition, we propose a progressive strategy for representation refinement of SRC. Secondly, to adaptively optimize progressions, a progressive index called Max Energy Residual Index (MERI) is defined as the guidance. Furthermore, we extend PSRC to bimodal biometrics and propose a Competitive PSRC (C-PSRC) fusion approach. The C-PSRC creates more discriminative fused sample and fusion dictionary by comparing residual errors of different modalities. By comparing with several state-of-the-art methods on three finger vein benchmarks, the superiority of the proposed PSRC and C-PSRC is clearly demonstrated.

A Survey of Face Recognition Methods Based on Mask Region Restoration Algorithm

The complex and diverse background conditions, the variability of human face and the artificial deliberate transformation have all become the problems that traditional face recognition technology can't solve, and they can't really reflect the imperceptible advantages of face recognition. After wearing masks, the most important mouth and nose areas of human face are covered by these masks with different shapes and colors, and some facial features are hidden, and the key feature points that can be extracted are greatly reduced. In this paper, the existing mask region repair algorithms at home and abroad are classified and summarized, mainly including mask occlusion face recognition based on robust occlusion, mask occlusion face recognition based on sparse representation classification and mask occlusion face recognition based on neural network. It is pointed out that partial occlusion is one of the main difficulties. The main methods and shortcomings of face recognition based on mask region repair algorithm are systematically analyzed and summarized, and the main problems and possible research ways in the future are analyzed.

A Sparse Representation Classification Scheme for the Recognition of Affective and Cognitive Brain Processes in Neuromarketing.

In this work, we propose a novel framework to recognize the cognitive and affective processes of the brain during neuromarketing-based stimuli using EEG signals. The most crucial component of our approach is the proposed classification algorithm that is based on a sparse representation classification scheme. The basic assumption of our approach is that EEG features from a cognitive or affective process lie on a linear subspace. Hence, a test brain signal can be represented as a linear (or weighted) combination of brain signals from all classes in the training set. The class membership of the brain signals is determined by adopting the Sparse Bayesian Framework with graph-based priors over the weights of linear combination. Furthermore, the classification rule is constructed by using the residuals of linear combination. The experiments on a publicly available neuromarketing EEG dataset demonstrate the usefulness of our approach. For the two classification tasks offered by the employed dataset, namely affective state recognition and cognitive state recognition, the proposed classification scheme manages to achieve a higher classification accuracy compared to the baseline and state-of-the art methods (more than 8% improvement in classification accuracy).

Active Learning for Hyperspectral Image Classification Using Kernel Sparse Representation Classifiers

Active learning is one of the popular approaches that can mitigate some of the drawbacks of supervised classification. Although sparse representation classifier has already proven to be a robust classifier and successfully used in many applications, it is seldom used jointly with active learning. In this letter, we propose a novel active learning technique for sparse representation classifiers. In the proposed model, the query function is designed by combining uncertainty and diversity criteria, both of which are defined by using the sparse representation classifier in kernel space. The proposed technique outperforms other state-of-the-art methods in terms of classification performance.

Cumulative foot pressure image recognition via Gabor filters and sparse representation classifier

Analysis of cumulative foot pressure images (CFPIs) can be utilised like other biometrics for personal recognition. This biometric, along with other biometrics such as gait, can be helpful in identification. Also, this overcomes some drawbacks of common biometrics (such as face and gait) in data recording and taking people's foot data without knowing them. In the process of capturing cumulative foot pressure images, the spatial and temporal changes of ground reaction force during one gait cycle are recorded. In this biometric, there are some challenges such as walking at different speeds. In this paper, we present a new approach based on Gabor filters as a feature vector and sparse representation classification (SRC). To reduce the feature dimension, eigenfoot and linear discriminative analysis (LDA) are also used. To obtain translation and rotation invariant representation, a normalisation preprocess is applied. We evaluate the proposed approach via dataset D of the Chinese Academy of Science (CASIA) Gait-Footprint dataset containing the cumulative foot pressure images of 88 persons. The experimental results indicate that the proposed method has higher accuracy than other tested methods, and also is robust to changes in walking speed.

Cumulative foot pressure image recognition via Gabor filters and sparse representation classifier

Cumulative foot pressure image recognition via Gabor filters and sparse representation classifier

Anticancer Peptides Classification Using Kernel Sparse Representation Classifier

Cancer is one of the most challenging diseases because of its complexity, variability, and diversity of causes. It has been one of the major research topics over the past decades, yet it is still poorly understood. To this end, multifaceted therapeutic frameworks are indispensable. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Anticancer peptides</i> (ACPs) are the most promising treatment option, but their large-scale identification and synthesis require reliable prediction methods, which is still a problem. In this paper, we present an intuitive classification strategy that differs from the traditional <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">black-box</i> method and is based on the well-known statistical theory of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">sparse-representation classification</i> (SRC). Specifically, we create over-complete dictionary matrices by embedding the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">composition of the K-spaced amino acid pairs</i> (CKSAAP). Unlike the traditional SRC frameworks, we use an efficient <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">matching pursuit</i> solver instead of the computationally expensive <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">basis pursuit</i> solver in this strategy. Furthermore, the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">kernel principal component analysis</i> (KPCA) is employed to cope with non-linearity and dimension reduction of the feature space whereas the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">synthetic minority oversampling technique</i> (SMOTE) is used to balance the dictionary. The proposed method is evaluated on two benchmark datasets for well-known statistical parameters and is found to outperform the existing methods. The results show the highest sensitivity with the most balanced accuracy, which might be beneficial in understanding structural and chemical aspects and developing new ACPs. The Google-Colab implementation of the proposed method is available on the GitHub page (https://github.com/ehtisham-Fazal/ACP-Kernel-SRC).