Abstract: A system and method for processing video data of an object in a movement space comprises uses a plurality of cameras each facing the movement space. An object detection engine running on a processor is in communication with each of the plurality of cameras that transmit video data of the movement space to the processor. The object detection engine determines the coordinates of a bounding box that closely surrounds the object in a frame of the video data from a master camera. The processor then determines, epipolar lines corresponding to points on the bounding box in the frame of the master camera in each of the frames of the video data from the non-master cameras. The processor crops the video data in the frame of the non-master cameras outside two of the epipolar lines to reduce and improve the video data processing.