Multimanifold analysis with adaptive neighborhood in DCT domain for face recognition using single sample per person

Appearance-based face recognition methods have achieved great success in face recognition, whereas these methods fail to work for face recognition from single sample per person (SSPP). However in the most real-world situations there is only one image per person available such as law enhancement, epassport and ID card identification. In this paper a novel mutimanifold learning techniqe called improved-DMMA (I-DMMA) is proposed to address the SSPP problem. I-DMMA, is an improved version of DMMA which automatically determines the local neighborhood size and utilizes discrete cosine transform (DCT) as an initial feature extraction step. Experimental results on a widely used face database FERET, is presented to demonstrate the efficacy of the proposed approach.