Abstract
Sign languages can be decomposed to four components including hand positions, movements, shapes, and orientations. This paper considers several features concerning hand movements for sign language recognition. Motion Chain Code (MCC) is one of the methods to represent a hand trajectory. Mahmoud et al. tried to recognize the hand movements through the MCC. It however had some troubles with depth perceptions, because they considered only 2D movements (up, left, down, and right). In this paper, the depth movements of hand are recognized on basis the wrist area changes in sign gestures. Recognition performance of the proposed MCC in corporating the depth movements is substantiated by some experiments. As a result, the recognition rate was improved by the MCC including depth information.
