“Accidents are common in sports”, is a statement that we hear often. The gravity of some mischances is massive that it might persist in the long run. The athletes and sports personnel are well adjusted to the idea of any possible mishap and they try to adapt with it. However, Chronic Traumatic Encephalopathy, most commonly known as CTE is not such an accident that occurs one fine day and can be treated in an instant. But it grows gradually and if not diagnosed and addressed in an early stage, has the capacity to become a fatal brain disease.

CTE’s are commonly found in the athletes of contact sports such as hockey, American football, boxing, wrestling, rugby etc. Difficulties that arise due to CTE are cognitive and emotion-related, physical as well as behavioural. In a study conducted on CTE, it was found that it is associated clinically with memory issues, personality changes, Parkinsonism, speech and gait abnormalities. The issues pertaining to CTE begins with blows occurring on the head.

It is to prevent accidents and the development of conditions similar to CTE, sportspersons are recommended to wear protective gear while on the field. High-quality protection has contributed to the prevention of similar occurrences to a certain extent. In a situation where the profundity of brain damages are grave, improving safety guards are the only measures to fight it. 

Artificial intelligence is stepping into the area and is doing its share for finding means to improve the protection of athletes. The National Football League- a professional football association in America, teamed up with Amazon Web Service (AWS) and came up with an AI challenge. The competition required scientists across the world to teach computers to automatically detect head impacts of players from NFL games footages. 

Computer Vision in detecting head impacts

The scientists at the AI challenge, came up with ground-baking CV models to strengthen the efforts of the NFL to understand possible head injuries among players. NFL reviews initially studied the injuries by analysing the footage of serious accidents. On a frame-by-frame basis, from different angles, by recording over 150 variables, they used to understand the possibility of a concussion or any such internal damage. The computer vision-based winner’s model makes this process comprehensive, accurate and 83 times faster than manual analysis. The conclusion derived from the models are used for possible innovations that can be made by improving safety measures, and coaching and training strategies. 

More than 1000 data analysts from 65 countries across the world participated in the challenge. Computer Vision, a popular AI-grounded technology that is becoming a major trendsetter of all the sectors in recent times, commonly detect 2D images. But these challenges required the minds to work on the 3D location of the players on the field which is a leapfrog in the application of CV in contact sports. 

Future of the models

The new models will be helping the analysts build a “Digital Athlete”, a virtual representation of a player through which they will be capable of predicting the injury. After each game, the data is inputted in the model, through which the Digital Athlete will be able to run N number of simulations and thereby study the impact on the health of the players. Through this, the Digital Athlete will help in the real-time risk analysis by documenting the position, speed acceleration and orientation of every player. Furthermore, it can identify as well as evaluate the quality of the safety measures adopted. 

Dr. Priya Ponnapalli, Senior Manager at the Amazon Machine Learning (ML) Solutions Lab, stated about this AI-based advancement that it, “bring us steps closer to our goal and I couldn’t be more thrilled to see how this work transforms the sport in the coming years”.