Segmentation Of Complex Gesture Based On Multi-Attributes

As a widely used communication method, dynamic gesture excels in message delivering. To better describe movement of gesture, the goal of achieving a concise and precise trajectory segmentation was put forward. In this article, analyze on complex gesture movement is presented, and a prime segmentation algorithm is fulfilled.Initially, by studying the three-dimensional features of gesture movement, multi-attributes is utilized to describe the trajectory. The integrity of this part is tremendously improved by the employ of torsion properties. Built on the completion of this part, a mono segmentation method with bandwidth is proposed. This method eliminates affects from ambiguity of gesture trajectory, which improves the reliability.Having considered the need for gesture recognition, two methods are proposed to segment the trajectory. The first is based on the principle of minimum description length, while the other is built on the reliability. In this article, we focus on the latter. Reliability is used to describe the character points under multi-attribute segmentation, followed by grouping the analogical segment points and using threshold filtering to obtain results. This method perfectly balances the accuracy and conciseness, which meets the need of our project, making it the main achievement of this article.A model based on three-dimensional gesture movement interaction is designed, following the human-computer interaction principles. Sophisticated gestures can be recognized by segmentation via this model, confirming the validity and availability of the idea of this article.