Will A.I. finally put an end to unfair judging in gymnastics?

Disputes over unfair scoring and judging in sports are as old as sports themselves are. In many sports like athletics, tennis and football, modern technologies, photography and film have been around for a while and have assisted line and finish line judges for decades. Goal-line technology, approved by FIFA in 2012, finally put an end to goal disputes at major football events where it’s been put to use. Unfortunately in some other sports, like figure skating and gymnastics, where skills are performed in split seconds and artistic presentation also affects the final score, the controversies are still ongoing and seem to plague major competitions, even high profile international ones like the Olympics. In some of these sports medals are lost and won over just a few degrees difference in the athlete’s rotation in jumps, bent elbows or knees, and wobbles or steps during landing. As history has proven over and over again, the human eye and mind are just not sensitive enough to make fair judgements, especially when politics and age old rivalries between nations come into play.

A new artificial intelligence based solution by the Japanese technology giant Fujitsu, named “Judging Support System” was recently approved by the International Gymnastics Federation (FIG) and it will be implemented at large international gymnastics competitions starting this year.

The Judging Support System’s algorithm is able to analyse quick movements and athlete performance at a level of accuracy that’s impossible for the naked human eye. Utilising Fujitsu’s MEMS solid-state Lidar sensors, it has a 15 meter detection rate and it generates 2 million pulses per second. The system’s key modules are movement sensing, joint position recognition and a database of gymnastics skill elements. A high-speed matching of captured data with previously stored data is the key to the system. Deep learning technology is utilised for the joint recognition module: the neural network model receives several multi-viewpoint depth images as input and outputs corresponding 3D joint position results. Scoring for both difficulty (D) and execution (E) can be aided by the Judging Support System. It currently contains a growing library of 800 possible elements for male gymnasts and just over 500 elements for female gymnasts.

Fujitsu’s marketing video for the Judging Support System

2019’s FIG World Cup Series will be used as a testing ground for the AI Judging Support System, with an official product launch scheduled in time for the 49th Artistic Gymnastics World Championships in Stuttgart, Germany in October. For next year’s Tokyo Olympics, the system is scheduled to support men’s vault, rings and pommel horse; as well as women’s vault and balance beam. The plan is to implement it for all gymnastics events (4 for women and 5 for men) by the Paris Olympics in 2024.

Solutions that rely on data science and artificial intelligence have been widely used in many other sports to evaluate players, determine game outcomes or even revisit historical decisions and solve decades old sports mysteries, like the truth behind the notorious goal in the 1966 World Cup Final. Computer vision, a field of artificial intelligence and computer science has been used to evaluate athletes in football, basketball, hockey and baseball. In fact it’s now used by teams to scout out players with high potential. “Everything that happens on the field, terabytes of data per game, is being collected, so this is where it’s really getting exciting now. In some cases it gives you insights you would never think of, and those instances can be game-changers.” Said Ari Kaplan, an American major league scout and baseball data analyst who has consulted for multiple Major League Baseball teams.

Now as AI is set to make scoring and judging more fair in gymnastics, more questions will be brought up about pressure on the athletes and the reaction of viewers and spectators. No change in scoring and judging can slip by without affecting the game and the players themselves. A controversial change in the calculation of gymnastics scores was made in 2006 with the implementation of the “open scoring system” that put more emphasis on the difficulty of skills performed and less on perfect execution. The latter resulted in “perfect tens” in the past that raised athletes like Nadia Comaneci to superstar status. The former forces athletes to continue raising the difficulty level of their routines, and according to some it has led to a significant increase in injuries.

How the Judging Support System, will change gymnastics remains to be seen. But we can be certain that it will make the judges’ work easier and maybe, just maybe it will minimise bias.

Many thanks to Istvan Turcsik, one of the Hungarian National Gymnastics Team’s coaches, for leading me to this story.

Bogi Szalacsi is a Senior Associate with infoNation, based in London. You can contact her at bogi@infonation.io and follow her on Twitter: @infoNation5.