Dorsal Hand Recognition Through Adaptive YCbCr Imaging Technique

摘要

Dorsal hand recognition is a trending topic in biometrics and human computer interactive systems. The characteristic and unique shape of the dorsal side of users' hands could be identified and discriminated for continuous authentication or could be tracked for second security option as a keyboard passwords. Therefore we propose a novel recognition system that deals with users' hands on the keyboard using adaptive YCbCr color space. The images are extracted from a video recorded by a camera mounted on the monitor and the Cb and the Cr color intervals of the dorsal hands are identified and stored. In contrast with the common algorithms that deal with the static interval, we propose an adaptive system which initially identifies the Cb and Cr values of the users' hands and subsequently recognize the dorsal hands throughout the frames of the video.