Object Recognition in Soccer Sports Videos

Object recognition plays a vital role in many applications in computer vision. In sports video recognition an action is a major part, in that object recognition is a key requirement. In the video, many challenges are faced which includes fast motion, occluded objects, different sizes of objects, difficult illumination, and continuous change in the background in identifying the object is a major task. The proposed system’s main aim is to deliver a summary of the existing system’s object detection approaches that belongs to CNN and debug their performance on soccer sport video and their training videos and match with the input soccer video. The performance of object recognition is discussed in various situations.