GreenSea: Visual Soccer Analysis Using Broad Learning System

Abstract

Modern soccer increasingly places trust in visual analysis and statistics rather than only relying on human experience. However, soccer is an extraordinarily complex game that there exist no widely-accepted quantitative analysis methods. And the statistics collection and visualization are time-consuming which result in numerous adjustments. To tackle this issue, we developed GreenSea, a visual-based assessment system designed for soccer game analysis, tactics, and training. The system uses broad learning system (BLS) to train the model, in order to avoid the time-consuming issue that traditional deep learning may suffer. Users are able to apply multiple views of a soccer game, and visual summarization of essential statistics using advanced visualization and animation are available. A marking system trained by BLS is designed to performance quantitative analysis. A novel recurrent discriminative broad learning system (RDBLS) is proposed to carry out a long-term tracking. In our RDBLS, the structure is adjusted to have better performance on binary classification problem of discriminative model. Several experiments are carried out to verify that our proposed RDBLS model can outperform the standard BLS and other methods. Two studies were conducted to verify the effectiveness of our GreenSea. The first study was on how GreenSea assists a youth training coach to assess each trainee's performance for selecting most potential players. The second study was on how GreenSea was used to help U20 Shanghai soccer team coaching staff analyze games and make tactics during the thirteenth National Games. Our studies have shown the usability of GreenSea and the values of our system to both amateur and expert users.

Digitalization and Information Management Mechanism of Sports Events Based on Cooperative Sensing Model of Multisensor Nodes

“…Many sensor nodes that can communicate by wireless communication are randomly deployed in the monitoring area. After the sensor nodes collect the data, the data is transmitted to the aggregation nodes through multihop routing, and then, the data is sent to the task management console through Internet transmission, and finally, the user can obtain the data through the control center for subsequent data analysis and operation, and the transmission network is managed and configured accordingly [17]. Generally, the functions of sensor nodes include sensing, collecting, forwarding, and processing information, and its capacity size, communication capability, and ability to process data are less than that of aggregation nodes, which can connect wireless sensor networks to external networks.…”

DNN-based multi-output model for predicting soccer team tactics

“…In studies that visualize soccer formations, real soccer formations such as 4-3-2-1 are visualized ( Wu et al, 2018 ). A study using video images of soccer visualized analyses such as team tactics and player evaluation ( Sheng et al, 2020 ). In a study based on pass data from soccer games, the changing patterns of tactics were visualized by comprehensively analyzing pass data, including the related dynamics ( Xie et al, 2020 ).…”

Multi-Agent Broad Reinforcement Learning for Intelligent Traffic Light Control

“…Recently, Stacked BLS has been proposed to enhance the accuracy rate of models [40]. Moreover, the BLS and its variants [41], [42], [43] have been used in other fields, including image classification [44], [45], industrial process [36], resource utilization [46], medical care [47], time series prediction [48], maximum information network [49], network traffic flow prediction [43], [50], event commentary [51] and as well as traffic forecasting [52]. In the traffic area, the BRL methods have been first adopted in TLC [21].…”