Disney Researchers Build Automated Cameras To Follow & Record A Basketball Game
The time has come when the world is looking for robots to further ease the lives of human beings. Recently, some Disney research scientists reported that they’ve developed a system that could automate the cameras to record a complete Basketball game. The robotic cameras they've developed are being trained to follow the game in a way any human operator would. However, a major concern for researchers is that the system is not yet comparable to a flesh-and-bone cameraman who covers the game with remarkable aplomb, especially because Basketball is a fast-paced game.
An engineer at Disney research Peter Carr and a PhD student Jianhui Chen worked on training the cameras to follow the game by anticipating where the ball is going rather than keeping an eye on its position. In a game like basketball, where the ball is constantly in motion, tracking ball could result in jerky action which makes it difficult to watch the game. Hence, there’s a lot of scope for improvement in the recently developed system. Usually players know their positions at different times and this is what gives clues about the ball’s location.
The basketball court is divided into four quadrants to track the motion of ball and create a map of players' positions. The automated cameras will gain information from these maps to predict the players' next location. The currently implemented algorithm can’t be of much use to judge the team-work while passing the ball on the court. However, research shows that it is getting better in predicting where the ball could be in the next move. Researchers presented their first findings at IEEE Winter Conference on Applications of Computer Vision. Though the system is tested only in a basketball game, Carr and Chen expect it be used in other sports as well.
Source: #-Link-Snipped-#
An engineer at Disney research Peter Carr and a PhD student Jianhui Chen worked on training the cameras to follow the game by anticipating where the ball is going rather than keeping an eye on its position. In a game like basketball, where the ball is constantly in motion, tracking ball could result in jerky action which makes it difficult to watch the game. Hence, there’s a lot of scope for improvement in the recently developed system. Usually players know their positions at different times and this is what gives clues about the ball’s location.
The basketball court is divided into four quadrants to track the motion of ball and create a map of players' positions. The automated cameras will gain information from these maps to predict the players' next location. The currently implemented algorithm can’t be of much use to judge the team-work while passing the ball on the court. However, research shows that it is getting better in predicting where the ball could be in the next move. Researchers presented their first findings at IEEE Winter Conference on Applications of Computer Vision. Though the system is tested only in a basketball game, Carr and Chen expect it be used in other sports as well.
Source: #-Link-Snipped-#
0