Sparse-Representation-Based Classification with Structure-Preserving Dimension Reduction

Abstract

Sparse-representation-based classification (SRC), which classifies data based on the sparse reconstruction error, has been a new technique in pattern recognition. However, the computation cost for sparse coding is heavy in real applications. In this paper, various dimension reduction methods are studied in the context of SRC to improve classification accuracy as well as reduce computational cost. A feature extraction method, i.e., principal component analysis, and feature selection methods, i.e., Laplacian score and Pearson correlation coefficient, are applied to the data preparation step to preserve the structure of data in the lower-dimensional space. Classification performance of SRC with structure-preserving dimension reduction (SRC-SPDR) is compared to classical classifiers such as k-nearest neighbors and support vector machines. Experimental tests with the UCI and face data sets demonstrate that SRC-SPDR is effective with relatively low computation cost

A Comparative Study of 2D UMI and 3D Zernike Shape Descriptor for ATS Drugs Identification

“…The proposed feature extraction method is going to be compared again with 3D Zernike discussed in this chapter and more benchmarking 3D-based SDs, such as 3D Legendre and 3D Fourier transform, will also be compared using specifically tailored classifiers for shape representation, such as classifiers by Zhiyong et al [46], Xu et al [47]. Additional data from National Poison Centre, Malaysia, will also be used as dataset in the future works.…”

Cascaded regression with sparsified feature covariance matrix for facial landmark detection

“…It is interesting to note in here that several works have tackled the problem of feature selection following a sparse-inducing criterion Zhang et al (2011);Xu et al (2014). Our work di↵ers from these in that we generate sparsity on the covariance matrix rather than on the original feature space.…”

Robust face recognition using sparse representation in LDA space

“…Recently, sparse representation paradigm [7] has been intensively investigated, demonstrating its effectiveness in several fields such as multi-class feature selection [50], image restoration [9], data compression [14], visual tracking [21,48], image classification [41,44] and, not the least, face recognition systems.…”