| CPC A63B 71/0605 (2013.01) [A63B 69/0068 (2013.01); G06T 7/248 (2017.01); G06T 7/66 (2017.01); A63B 2220/05 (2013.01); A63B 2220/806 (2013.01); G06T 2207/10016 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30196 (2013.01)] | 18 Claims |

|
1. A method for scoring performance of a performer in a sport, the method comprising:
receiving a video feed of a performance, the performance is by the performer;
recognizing, by computer vision, predefined body parts of the performer in frames of the video feed;
determining a geometric center of each body part of the predefined body parts;
building a framework by connecting the geometric centers of the predefined body parts by projecting a body of the performer in a 2D or 3D space;
tracking spatiotemporal changes in the framework to determine predefined body movements and positions, wherein the spatiotemporal change in the framework is proportional to movements of body parts in relation to each other; and
scoring the performance, by a scoring module, wherein the scoring module is based on a deep neural network, the scoring module trained using established rules and standards for scoring.
|