A Face Recognition Approach Based on Entropy Estimate of the Nonlinear DCT Features in the Logarithm Domain Together with Kernel Entropy Component Analysis
Arindam Kar, Debotosh Bhattacharjee, Dipak Kumar Basu, Mita Nasipuri, Mahantapas Kundu · International Journal of Information Technology and Computer Science · 2013
This paper exp loits the feature ext raction capabilit ies of the discrete cosine transform (DCT) together with an illu mination normalization approach in the logarithm domain that increase its robustness to variations in facial geo metry and illu mination.Secondly in the same domain the entropy measures are applied on the DCT coefficients so that maximu m entropy preserving pixels can be extracted as the feature vector.Thus the informative features of a face can be extracted in a low dimensional space.Finally, the kernel entropy component analysis (KECA) with an extension of arc cosine kernels is applied on the ext racted DCT coefficients that contribute most to the entropy estimate to obtain only those real kernel ECA eigenvectors that are associated with eigenvalues having high positive entropy contribution.The resulting system was successfully tested on real image sequences and is robust to significant partial occlusion and illu mination changes, validated with the experiments on the FERET, AR, FRA V2D and ORL face databases.Experimental comparison is demonstrated to prove the superiority of the proposed approach in respect to recognition accuracy.Using specificity and sensitivity we find that the best is achieved when Renyi entropy is applied on the DCT coefficients.Extensive experimental comparison is demonstrated to prove the superiority of the proposed approach in respect to recognition accuracy.Moreover, the proposed approach is very simp le, co mputationally fast and can be imp lemented in any real-time face recognition system.