Discriminative GoDec+ for Classification
Kailing Guo, Xiangmin Xu, Dacheng Tao · IEEE Transactions on Signal Processing · 2017
GoDec+ is a robust low-rank representation model, which adopts correntropy to model noise and corruptions. To extend GoDec+ for classification, this paper proposes discriminative GoDec+ (D-GoDec+). In the model, each class is represented by a shared underlying subspace and a specific transformation matrix. Structural label information and the Fisher discrimination criterion are incorporated to model the reconstruction errors and coefficients. An efficient solution to D-GoDec+ is proposed based on half-quadratic optimization, and convergence of the solution is rigorously analyzed. Based on D-GoDec+, a simple but effective classification method is presented by combining the discriminability of reconstruction errors and coefficients. Through the use of transformation matrices, the classification method avoids complex encoding computation in the dictionary represented methods and thus is very efficient. Experimental results on face recognition, object classification, scene classification, and action recognition demonstrate the advantages of the proposed model.