Gomoku piece recognition method based on improved Yolov8

To recognize Gomoku piece types, based on the enhanced YOLOv8, a Gomoku piece method was proposed. First, the relevant photos of the Gomoku pieces were collected and made into a dataset. Then, the concept of Focal Loss is integrated into the original model to make it have higher recognition accuracy and less loss. Finally, the improved model is trained to obtain a model that can accurately identify the types of chess pieces. The experimental findings indicate that the proposed model demonstrates strong performance across various metrics, including accuracy, recall rate, and mean average precision, while also achieving a significant reduction in loss value. At the same time, the type of chess pieces in various cases is accurately judged, which meets the needs of chess pieces recognition. In a word, the Gomoku piece recognition method proposed is accurate in judging the type of pieces and provides a reference method for the task of recognizing the types of pieces, which holds substantial value for visual identification in gaming robotics.