Research On The Algorithm Of Hand Shape Feature Extraction Of Biological Recognition Technique

With the development of social economy, biometrics characteristic recognition received more and more attentions. It is an efficacious method for recognizing person's legal identity. Because of the biometrics of humans have uniqueness and invariability, that outstanding advantages are that other conventional authentication methods lack.Because of the simple collection instrument (compared with fingerprint and iris) and the high verification speed, hand shape verification becomes another method of biometrics characteristic recognition following the fingerprint verification. Reports show that hand shape verification has been widely used in access control, time and attendance, and personal identification fields throughout the world during the past several years. In the foreign biometrics characteristic recognition commodity market, the hand shape verification systems have shared the half with fingerprint systems. But techniques of the hand shape recognition and verification in our country are still in the primary development phase.Authentication algorithms of hand shape are researched in this thesis by using PC and MATALB7.0 software. This dissertation present a hand shape verification technology as left and right half axis width for feature. That is optimized the algorithm of authentication as finger's width for feature. Compensate the erroneous judgment which was caused by exceptional circumstance; Avoid the activities of non-linear error of the fingers by separating the fingers; Re-use of a coordinate system, as a result of image offset rotation, translation, and the impact of expansion. Compared with characteristic vectors matching method and the point pattern matching method. The placing of the fingers will have more flexibility by cancel the fixing pin fixed on the fingers. Because of the left and right half axis width were computed by point, so that the algorithm has the information of point feature which can guarantee the high computational speed, balanced the wrong reacceptance and the wrong reject rate.