Hand vein recognition with rotation feature matching based on fuzzy algorithm

The Bodily motion or emotion, which can be obtained for example from a hand or a face, originates gestures. Every individual has a unique pattern of dorsal hand veins. The vein pattern's orientation changes when one rotates their hand in a particular direction. This study focused on hand-gesture recognition using dorsal hand veins. The aim of this work is a novel technique to track and recognizing hand vein rotation using fuzzy neural network, and the change in orientation was considered as a gesture and measured.  The algorithms were tested over various rotations ranging from -۴۵^{\circ} to +۴۵^{\circ}. We successfully detected various rotations in both clockwise and anti-clockwise directions, achieving ۹۳\% accuracy and a reasonable time execution. This problem can be solved because a person can steer a car wheel merely by rotating his/her hand. An infrared camera captured the rotation of hand veins, so car wheel steering was unnecessary.

Keywords:

Complex Walsh transform , Dorsal hand vein pattern , feature extraction , Fuzzy neural network , Sectorization