At the 2024 Paris Olympics, which is in full swing, AI is frequently seen in various fields.
Diving stars Quan Hongchan and Chen Yuxi use Baidu's AI diving auxiliary training system in their daily training to carefully refine their movements and accurately grasp the timing of opening their bodies and entering the water.
The Chinese National Sailing Team uses Lenovo AIPC's built-in personal intelligent agent "Xiaotian" to develop exclusive training plans and analyze body movements and movement trajectories to improve training results.
Not only that, technologies such as big data analysis and virtual reality (VR) have long been integrated into the training and rehabilitation stages of athletes, providing strong support for improving their competitive level. At the same time, it covers sports performance analysis, game prediction, intelligent referee assistance, athlete health monitoring, tactical analysis, virtual training and simulated games and other aspects.
During the Paris Olympics, the World Anti-Doping Agency (WADA) also used AI technology to conduct anti-doping tests to ensure the fairness of the competition. AI-driven Alibaba Cloud intelligent stroboscopic analysis technology has been used in refereeing work, especially in diving competitions, where AI cameras can capture athletes' movements in real time and assist referees in making more accurate decisions. Another example is Dingqi Sports Products in Huai'an, Jiangsu, which implants smart chips in the inner liner of a football to record the ball touch at a frequency of 500 times per second, providing accurate data support for the video assistant referee (VAR) system.
The broadcasting of the Paris Olympics also ushered in major changes. As the official cloud service partner of the Olympics, Alibaba Cloud has achieved a transition from satellite broadcasting to cloud broadcasting, using new AI-enhanced broadcasting technology, combined with multi-lens playback system, multi-camera shooting and 3D reconstruction technology to give viewers a new viewing experience. For the first time, the Paris Olympics adopted an AI-driven slow-motion playback camera system provided by Alibaba Cloud, which will be deployed in 14 venues covering 21 sports. For example, innovative technologies such as "bullet time" playback and 3D freeze-frame images allow viewers to appreciate the details of the event from different angles and in all directions, as if time has stopped.
Another Chinese company is also using AI to empower event broadcasting. In July 2024, SenseTime and Shanghai Media Technology Co., Ltd. (SMT) jointly developed the "InnoMotion Smart Event Broadcasting Solution", which mainly provides full event broadcasting for table tennis and archery. Based on its own visual big model system, SenseTime has developed a multi-camera visual motion capture system with functions such as intelligent highlight moment capture, wonderful moment generation, technical data statistics, motion analysis and judgment, and event live broadcast support to assist athletes in improving their performance. This event broadcasting solution uses 3D non-perceptual motion capture technology to achieve multi-scene coverage of multiple people, large ranges, and multiple angles without wearable devices, and obtain spatial motion posture information in real time, especially for accurate capture and real-time analysis of high-speed small targets such as table tennis and archery arrows.
"The real-time trajectory of table tennis ball speed and the restoration of arrow trajectory in archery are the first in the broadcast of international events including the Olympics," Lin Yiting, director of marketing and solutions at SenseTime's overseas R&D center, told Jiemian News. In the table tennis event, SenseTime's self-developed non-perceptual 3D video motion capture technology can collect the speed and direction of the ball in real time. At the same time, combined with the SMT table tennis speed and direction AR visualization system and the broadcast vehicle slow motion playback system, the visual presentation of the speed and direction of the ball is realized.
Secondly, AI technology is used to collect the data of the landing point and trajectory of table tennis balls in real time, and real-time classification and statistics are performed in the background, realizing data visualization in various forms. This solution not only enriches the data types of table tennis events, but also reduces the cost of broadcasting by providing lightweight deployment solutions and domestically developed algorithms. "The advance arrangement of equipment installation and emergency plan preparation at the Olympic Games is a very big challenge. This is also the first time we have broadcast the Olympic Games. There are some gray areas in the actual situation and routing requirements of overseas venues. We must be fully prepared to deal with various problems and changes that may arise." Lin Yiting said.
In addition, after the cameras and acquisition equipment were deployed on site, they needed to be debugged and deployed to optimize the online services. To ensure the accuracy of the online system in a short period of time was a huge challenge for the front-end and back-end support teams. In the end, the team successfully completed the preliminary preparations and debugging on site.
During the Olympics, SenseTime's big model also provided a variety of interactive experiences for ordinary audiences, such as Olympic-themed fun photo templates, which can generate their own Olympic-style photos with one click, as well as intelligent generation of personalized Olympic viewing calendars and Olympic data analysis and other interactive activities, allowing audiences to participate in the Olympics.
"AI+Sports is a great potential direction for the development of big models in the future." Xu Li, Chairman and CEO of SenseTime, said: The creativity of generative artificial intelligence has brought more interesting content creation and popularization methods to sports.
In professional sports training, the 3D motion capture system is used to analyze athletes in real time, realize data interaction and assist training. Applying AI technology to data statistics and game video processing can perform technical and tactical analysis, form pre- and post-game data reports, and slice the video to accurately analyze the performance of each athlete, helping to improve training efficiency and competitive level. Lenovo, which is actively betting on AI, is also involved. In the pre-match training of the Chinese National Sailing Team, the personal intelligent body "Xiaotian" built into the ThinkPad T14pAI Yuanqi will assist the coach in formulating a dedicated training plan during the training preparation stage. During the training process, it will use heterogeneous AI computing power to output analysis of body movements and movement trajectories to help the team fully grasp the training situation.
In the view of Thomas Bach, President of the International Olympic Committee, artificial intelligence provides athletes with personalized training methods and better sports equipment, greatly improves the organizational efficiency of sports events, changes the way sports are broadcast, and makes the audience experience more personalized and immersive.
This Paris Olympics has sent out a positive signal that "AI is reshaping sports events", and the entire sports industry may usher in unprecedented opportunities and changes.