부산대학교 도서관

A proposed first approach of a novel and simple biometric verification system based on the texture of the hand knuckles. A knuckle image with scale, rotation, and translation invariance is isolated from the hand recorded image with digital image processing techniques. The wrinkle of the knuckle images are extracted to a black and white image which is used as biometric features. A hidden Markov model and a support vector machine are proposed as verifiers. A similar equal error We of 0.094 is reached by both classifiers with our preliminary database consisting of 8 samples of 20 people's hands.