German researchers use neural networks to identify tennis players' emotions

TapTechNews June 21st news, according to the report of ScienceDaily on the 17th, researchers from the Karlsruhe Institute of Technology and the University of Duisburg-Essen in Germany accurately identified the emotions expressed by the body language of tennis players during the game with the help of computer-aided neural networks.

The team trained this AI-based model for the first time using actual game data, and the research results were published in the latest issue of the academic journal in the field of artificial intelligence, Knowledge Systems. TapTechNews attached a link: https://www.sciencedirect.com/science/article/pii/S0950705124004908

It is reported that researchers in sports science, software development and computer science from the two schools developed a special AI model to identify the emotional state of tennis players using convolutional neural networks, and analyzed the videos of tennis players in actual games using a pattern recognition program.

Professor Darko Jekauc from the Institute of Sports and Exercise Sciences at the Karlsruhe Institute of Technology said, "Our model can identify emotional states with an accuracy rate of up to 68.9%, which is even better than that of human observers and earlier automated methods."

The project team used real scenes instead of simulated or artificial scenes to train its AI system, which is an "important and unique" feature of this study. The researchers recorded a video sequence of 15 tennis players in a specific scenario, focusing on the body language shown when scoring or losing points. The video shows that the cues of the players include lowering the head, raising the arms to celebrate, the racket dropping or changing the pace, which can be used to identify the emotional state of the players.

After obtaining the above data, the AI will "learn" to associate body language signals with different emotional responses, and judge whether to win or lose a point according to the positive or negative of the body language.

In terms of specific application, the team said that this research can be used later for improving training methods, team dynamics and performance, as well as preventing burnout, and can also be used in other fields - including healthcare, education, customer service and automotive safety, etc.