Fujitsu’s skeleton recognition AI technology is being used by the Japan Skating Federation to support figure skating training. This AI helps digitize and analyze 3D human movement, making it easier to study athletes’ performances.
Traditional motion capture methods are difficult to use in training because they take a long time to set up and offer limited analysis. Even video-based marker less systems often struggle to accurately track fast or complex movements like jumps and spins. Fujitsu’s AI, originally developed for gymnastics, offers a more accurate way to study posture and movement in real-time, making it more useful in fast-paced sports like figure skating.
The system uses correction algorithms to reduce errors in posture recognition and photorealistic simulations to quickly generate large training datasets. This reduces the need for long manual processes.
Fujitsu plans to use this AI beyond sports. Possible applications include tracking worker movements in factories, helping with early detection of illness, and supporting animation or gaming in the entertainment industry. This fits into the company’s wider goal of using technology to address broader social and business challenges.
Leave a comment