Deep Learning Approach to Automated Data Collection and Processing of Video Surveillance in Sport Activity Prediction
Xiaowei Tang, Fang Li, V. Sakthivel and S. Kirubakaran
The improvements in sports platforms, modern advancements, and the proliferation of sports monitoring activities have taken the modern era to the comprehensive extension of sports video frames’ data collection and processing. Since the video processing techniques require image processing procedures, it has been observed that the demanding factors include optimization, efficiency, noise factor reduction in sports moving video processing. Hence in this paper, A Deep Learning Video Surveillance (DLVS) Technique for Automatic Data Collection and Processing of Sport Activity Prediction is proposed to improve the efficient factors by minimizing the time and data storage consumption. The significant activity prediction of sports frames is being achieved in an optimized manner. The detailed data and feature extraction in a moving sports video are optimized by introducing the Optimal Moving Frame Analysis Algorithm (OMFAA). Likewise, a data collection and processing protocol (DSPP) is presented to ensure effective pixel processing and classification of video frames in sports activity prediction, which is considered a significant factor in DLVS-ADCP. The simulation research is performed with an available sports video database, and it is validated based on the parameters like accuracy, performance, optimization factor, and reliability.
Keywords: video surveillance, sports activity, deep learning, frames collection