China's Olympic delegation has been given a high-tech boost at the ongoing Paris Olympics as artificial intelligence (AI) technology has been integrated into the team's athletic performance analysis systems, providing a new dimension to sports strategies as they aim for glory in the French capital.
This cutting-edge approach has been spearheaded by a dedicated team at the General Administration of Sport's key laboratory for sports tactical analysis at Shanghai University of Sport, offering technological and tactical support to the Chinese 3x3 basketball players.
During the Olympics, the team utilizes an AI basketball performance analysis platform, employing AI models for data statistics and game video processing. This provides crucial insights to enable the team to refine their training and form competitive game plans.
The team operates in two sections: the front-end developers capture live videos, while the back-end analyzes this content to swiftly offer suggestions and strategies for upcoming games.
"We have built a 3D video environment in the stadium, and then used our algorithm to reconstruct the 3D model to quickly analyze the athletic and physical performance of our players. Then in the space of around 30 minutes to an hour, our data will be available and we can feed it back to our teams, athletes and coaches. In real time, you get to see the performance of the athletes, including the strain they are under load, their movement and their speed. Especially for the 3x3 basketball which may have multiple games in one day, our team can provide targeted assistance and strategic advice on the recovery and adjustment as well as tactical arrangements for subsequent games or the game the next day," said Zhang Mingxin, head of the basketball team under the university's School of Athletic Performance.
The adoption of AI technology in this pivotal role and its ability to collect and analyze athletes' intensity without the need for athlete's to use wearable sensors helps offer more precise data for decision-making while enabling the competitors to concentrate on their game unhindered.
"Through computer vision and 3D motion capture, we've realized recording the intensity of an athlete's running in 3D space scenes without wearing certain gadgets. For example, by measuring the direction, distance, acceleration, and displacement of their running, we can calculate an athlete's real-time intensity in the game through these formulas. Through such a method, we can actually make corresponding analysis to benefit the opponents. So after a game, we can not only see our own intensity, but also see the potential differences between our opponents and us," said Cheng Peng, an executive of a scientific and technical corporation.
The AI-driven video analysis, combined with unique algorithms, provides technical support for athletes and coaches, marking a significant trend in sports development. "Traditional sports performance analysis or technical and tactical analysis has the disadvantage of having a long time to receive feedback. However, with the support of AI technology, we can not only greatly enhance the efficiency of our feedback, but also, based on the end-to-end characteristics of AI, we may even discover some problems in sports performance or tactical aspects that we have previously overlooked. Therefore, this also raises new demands on our talent training," said Wang Ran, deputy dean of the School of Athletic Performance.