Palmprint recognition using novel oriented gabor gradients.

In biometrics, feature extraction is an important step to extract the unique information from physiological or behavioural characteristics of individual such as palmprint, faceprint, speech and gait. In this paper, a novel feature extraction technique is proposed for person authentication using palmprint based on gradients of gaborized image called Oriented Gabor Gradients (OGG). To validate the proposed feature extraction method, palmprint recognition has been tested on both left and right palm of IITD database of 230 persons, PolyU palmprint database of 386 persons. The proposed OGG method is compared with histograms of oriented gradients (HOG), gabor transform, gaussian membership based features (GMF), absolute average deviation (AAD) and mean features. Experimental results show primacy of the proposed technique over the existing ones in the literature and achieved higher accuracy. Lastly, K-nearest neighbor is used to validate the matching stage. [ABSTRACT FROM AUTHOR]