Fujitsu Human Motion Analysis technology increases fairness in judging across competition for all ten apparatuses
Lausanne / Tokyo, 6 October 2023 - Fujitsu Limited and the International Gymnastics Federation (FIG) are introducing the Judging Support System (JSS), based on AI technology, following the 52nd World Gymnastics Championships in Antwerp. The joint development of the judging system for all 10 apparatuses is an important milestone in the partners' ongoing cooperation.
The advanced AI solution, which uses camera image analysis, provides greater accuracy, fairness and comprehensive support for performance judging at the World Gymnastics Championships. Fujitsu's AI technology improves processing speed and allows the complex movements of the gymnasts to be captured almost simultaneously. This allows judges to analyse competitors' performances with more precision and detail. The new judging support replaces the previous JSS system, which was sensor-based.
The introduction of the improved competition system coincides with Fujitsu's plans to launch its Human Motion Analytics platform, which optimises gymnastics training, in April 2024. The innovation is part of Fujitsu Uvance, which provides technology solutions for business and societal challenges. The new solution combines advanced AI technology and image analytics expertise gained through the JSS project, aiming to address many challenges in sectors such as healthcare, retail, manufacturing and entertainment.
Major milestone for Fujitsu
The increasing complexity and sophistication of both gymnasts' training, skills and equipment brings challenges. This is why the FIG and Fujitsu launched their collaboration in 2017 to improve the integrity of sport by solving challenges in judging and ensuring fairness and transparency. Use of the system started in 2019 at FIG competitions, including the World Gymnastics Championships.
The Judging Support System is available to FIG member federations and can be used as part of their training programmes. For example, it assesses the athlete's skills and mastery of elements during training periods. The cooperation partners also provide TV broadcasters worldwide with analytical data and visual enhancements based on system data. Moreover, the FIG will use the JSS in November 2023 to create new assessment guidelines using numerical values. These include joint angles and gymnasts' body positioning.
Annex: Human motion analytics
Fujitsu's Human Motion Analytics (HMA) is a data analysis platform based on the world's most advanced high-performance posture recognition technology, developed through the Judging Support System, adopted by the FIG. The technology uses a unique correction algorithm that significantly reduces the jitter (estimation error), which has been a persistent challenge in image analysis, of pose recognition through deep learning technology. This makes it possible to generate large amounts of training data for different environments, taking into account factors such as venue brightness, venue colours and equipment. The technology also features so-called 4D-Capture technology, which analyses human movements in four dimensions (three dimensions and time) and recognises actions as a series of connected movements rather than a snapshot of a pose. With these innovations, even complex and fast movements, common in gymnastics, can be consistently identified with high accuracy.
Fujitsu has independently defined about 100 basic movements to enable rapid implementation of these technologies to various manufacturing and retail locations.
Ultimately, Fujitsu aims to ensure that the technology developed for HMA not only contributes to the optimisation of different areas, but enables its users to address societal challenges through creativity and innovation.
Growing opportunities with real-world practical examples
Besides possible applications in the sports world, HMA offers potential for a variety of areas, such as entertainment, the development of preventive medicine and exercise therapy in healthcare, applications in improving customer service experience in retail and improving workplace training and knowledge sharing, as well as safety for workers in the manufacturing sector.
Realisation of new video experiences using 3D data from athletes, improvement of movement in form assessment, possible application in game and animation production.
Collaborate with experts in visualising correct movements for the development of preventive care and exercise therapy, and to support users in continuing exercises.
Retail and Manufacturing
Using ergonomics and other video analysis to improve safety, reduce worker strain and improve training and knowledge sharing for employees.