“In athletics, 100 meters is just 100 meters. Sometimes in gymnastics, it’s 95 meters or 105 meters,” said Olympic Coach Yuliy Kuksenkov, after his athlete lost her medal due to a change in scoring. Indeed, gymnastics is among the most subjective of sports, with everything hinging upon the opinion of human judges. However, a new AI system promises to introduce an unprecedented objectivity to gymnastics scoring by using motion tracking technology to precisely gauge the precision of skills.
The International Gymnastics Federation (FIG) is planning to introduce a robotic judge that uses a 3D sensory system to instantaneously evaluate and score routines, according to The Guardian. The system, developed by Japan’s Fujitsu, will be deployed at the 2020 Olympic Games in Tokyo. By using AI motion tracking, the product will aid in scoring and deliver other benefits for gymnastic athletes, coaches and audiences.
The Fujitsu system can break down and identify individual gymnastic skills, discerning the flips from the flops and the Yurchenkos from the Tsukaharas. It also can evaluate specific moves, such as determining the degrees of leg separation on a split handstand on the floor exercise or the duration of an iron cross on the rings.
The use of AI motion tracking may help iron out some of the inherent inconsistencies in gymnastic judging.
For example, human judges often subject athletes who perform early in the rotation to stricter scoring criteria than they do to competitors who compete later. This relaxing of standards is a result of the judges losing concentration on the small details because of the repetitive nature of observing and evaluating gymnastic routines.
“A judge must work for eight hours per day – does that allow the mental capacity to remain coherent? It’s not possible to maintain a coherent mind of criteria. Only the computer does,” said former FIG president Bruno Grande in a quote in The Guardian.
Furthermore, gymnastics judging is sometimes accused of bias based on political or personal motivations. The use of an impersonal AI technology could help mitigate or eliminate such charges.
Fujitsu’s system could bring benefits for gymnastics far beyond just judging.
The technology also could be applied to coaching, observing gymnasts’ practice sessions and providing feedback on the precision of their moves.
It also can help with television or internet coverage of gymnastic events, showing audiences the name of each move that comprises a routine. Such information could be presented in real time, as part of slow-motion replay.
Nirel Marofsky is project analyst for the cognitive engine and application ecosystem at Veritone. She acts as a liaison to strategic partners, integrating developers and their capabilities into the Veritone aiWARE platform. Learn more about our platform and join the Veritone developer ecosystem today.