WITHDRAWN: Time Series Analysis of Volleyball Spiking Posture Based on Quality-Guided Cyclic Neural Network

Abstract

Abstract Human body posture analysis is an important research topic in the field of artificial intelligence and computer vision. Volleyball poses are smaller and typical examples are more complex than human postures. Studying the time series analysis of postures in volleyball spiking is of great significance for analyzing the direction of the spike. Based on the analysis of volleyball spiking posture characteristics and the circulatory neural network algorithm, this paper studies the key technology applicable to the timing analysis of queued motion posture based on the continuity law of volleyball spiking postures - a new type of neural circulation network based on quality guidance .In order to solve the problem of ambiguity and accuracy of queuing motion posture, a new neural network algorithm combined with quality guidance is used to realize the time series analysis of typical volleyball spiking posture based on body local region tracking results. The study of the queuing movement posture part is to track the athlete's body part, introduce a new quality map method, extract the behavior characteristics of the limb, define the posture template, and learn the gesture template using the improved recurrent neural network. First, define a continuous dynamic motion pose template, and use the cyclic neural network to learn the determined volleyball pose template. Each group of independent volleyball spiking posture templates is learned to form a time series analysis of volleyball spiking postures. The advantages of the new quality map proposed in this paper over other common quality maps are verified by experiments, which are better than other quality maps in phase consistency and stability. Through subsequent verification, it is also proved that the improved RNN network proposed in this paper can accurately realize the time series analysis of volleyball spiking posture.