Video Object Extraction by Using Background Subtraction Techniques for Sports Applications

Segmenting out foreground object from its background is an interesting and important research problem in the video based applications. It has more importance in the field of computer vision due to its applications such as sports, security systems, video surveillance, etc. The background subtraction algorithms are used to analyze the player‟s activity in sports, to improve the performance of player by detecting the motion of the players in video sequences. The various algorithms like frame difference, approximate median, mixture of gaussian are compared and analyzed with real time sports videos. Mixture of Gaussian turns out to be best in reliability of extraction of moving objects, robust to noise, whereas the conventional algorithms result in noise and poor extraction of objects. The parametric analyzes of metric such as recall, precision, etc., gives the complete behavior of player.