An Approach to Dynamic Gesture Recognition Based on Instantaneous Posture

Abstract

Dynamic gesture recognition is a challenging problem. Existing algorithms often input gesture data indiscriminately for recognition, which further increases the difficulty of recognition and reduces the accuracy and efficiency of recognition. In this paper, a dynamic gesture recognition method is proposed by searching the effective instantaneous posture in dynamic gestures. To a certain extent, the invalid gesture data is filtered, and an effective gesture feature description is established for real-time gesture recognition. Through the design of comparative experiments, the proposed algorithm is verified to be better than the sliding time window algorithm in recognition accuracy and efficiency.