질감정보를 이용한 동적 손동작 인식

Abstract

In this paper, the recognition of dynamic hand gestures using texture information is proposed. In order to effectively recognize dynamic hand gestures, DSTW(Dynamic Space Time Warping) algorithm is utilized for searching the most similar path between the pre-defined models and candidate regions of hand. The existing method using DSTW algorithm can not exactly recognize the models because the path can be generated from the candidates including non-hand regions. Therefore, in this paper, a texture information is combined to identify only the paths generated from hand regions. As experimental results, the proposed method has improved the recognition rate, compared with the existing method.