BallJames

BallJames is fully automatic system of computer vision which generates data in the 3D format without the need for connection of sensors to players or a ball.

According to developers, BallJames automatically generates 3D-this of video images of football matches. At stadium fourteen cameras which write all movements in the field are mounted. Then BallJames generates own data, such as accuracy, the direction and speed of passing, run, force of jumps, the movement of the player, nonverbal communications between players and how close the ball contacts to a leg after the first contact.^[1]

The solution allows to trace and to objectively estimate actions and efficiency of each certain player and a command in general. The acquired information can be used in the most different purposes — for demonstration of a game from any foreshortening in virtual reality, for increase in efficiency of sports rates, for the purposes of coaching staff, for example, that the trainer could understand, the player was how tired and when to do replacement, for the choice of optimal team's line-up under the specific rival, etc.

Movements of players are modelled using algorithms of machine learning which are improved at task performance due to new experience. In the field a system appropriates a certain value to each action: to corner kick, a penalty, I graze, etc. Eventually these values change depending on their success. For example, the goal has high value, but pass which could have low value earlier can become more valuable as the platform masters a game.^[2]

Notes