ABSTRACT

Face recognition has attracted attention of the researchers. Face recognition becomes challenging if various factors are considered such as varying illumination, pose, facial expression and somewhat occlusion. The face recognition becomes more challenging if the single sample per person is available. In this paper a fusion of method is proposed to deal with single sample per person. Gabor transform is good for eliminating the orientation differences. An efficient ridgelet transform is proposed which effectively collects the meaningful rotational features. The results obtained from these two transforms are combined to classify the face image using support vector machine and distance based classifier. Experiments on FEI, JAFFE, ORL, UMIST, MUCT face datasets shows that the proposed method improves the performance in the scenario of one training sample per person.

Title	:	Recognizing faces with single sample per subject using fusion of transforms
Authors	:	Sujata G. Bhele, Vijay H. Mankar
Keywords	:	Gabor Wavelet, Feature extraction, Support Vector Machine
Issue Date	:	July 2016.
Abstract	:	iv>
Page(s)	:	221-228
ISSN	:	0975–3397
Source	:	Vol. 8, Issue.07