| Abstract [eng] |
This research focuses on the development of a system capable of detecting a basketball backboard, rim, net, and ball from a single static video stream. The system operates by processing frames extracted from the video, with the aim of achieving accurate and efficient object detection in real time. Different versions and configurations of YOLO models were tested and compared based on precision, inference speed, and model size. The ultimate objective is to identify the most effective setup to support a reliable shot recognition system. |
