Article: Gomoku game recognition and localisation using image processing and deep learning Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.9 pp.43 - 63 Abstract: Aiming at the accuracy of Gomoku game recognition during human-machine game, this study introduces a method for Gomoku game recognition and localisation using image processing combined with deep learning techniques, providing accurate position information for subsequent game decision-making and control driving in human-machine game devices. Firstly, the image processing technology is used to extract and correct the Gomoku board area, and then the corner points of the board are obtained by corner division. Secondly, to enhance the accuracy of Gomoku piece recognition, spatial attention (SA) is introduced into the C2f module of the backbone network of the yolov8 model to achieve the piece recognition. Finally, the location of the Gomoku piece is completed by comparing the position of the piece and the corner point. After experimental validation, the improved model shows a marked improvement over the original yolov8 model, with the precision rate, recall rate, mAP50 and mAP50-95 being 99.9%, 99.0%, 99.4% and 96.4%, respectively. Even under challenging conditions such as uneven brightness, low brightness and tilting of the camera, the recognition and position judgment of the Gomoku pieces can still maintain high accuracy. The introduced method has high performance in Gomoku piece recognition and localisation, providing favourable support for visual detection of Gomoku gaming devices. Inderscience Publishers - linking academia, business and industry through research

Title: Gomoku game recognition and localisation using image processing and deep learning

Authors: Huanxi Chen; Wei Li; Yong Shen; Xiang Ding; Shunbin Li; Minghao Gu; Dexin Sun; Zhenhua Tan; Guochao Wang

Addresses: School of Machinery and Transportation, Southwest Forestry University, Kunming, 650224, China ' School of Machinery and Transportation, Southwest Forestry University, Kunming, 650224, China ' School of Electrical Information Engineering, Yunnan Minzu University, Kunming, 650504, China ' School of Machinery and Transportation, Southwest Forestry University, Kunming, 650224, China ' School of Machinery and Transportation, Southwest Forestry University, Kunming, 650224, China ' School of Machinery and Transportation, Southwest Forestry University, Kunming, 650224, China ' School of Machinery and Transportation, Southwest Forestry University, Kunming, 650224, China ' School of Machinery and Transportation, Southwest Forestry University, Kunming, 650224, China ' School of Machinery and Transportation, Southwest Forestry University, Kunming, 650224, China

Abstract: Aiming at the accuracy of Gomoku game recognition during human-machine game, this study introduces a method for Gomoku game recognition and localisation using image processing combined with deep learning techniques, providing accurate position information for subsequent game decision-making and control driving in human-machine game devices. Firstly, the image processing technology is used to extract and correct the Gomoku board area, and then the corner points of the board are obtained by corner division. Secondly, to enhance the accuracy of Gomoku piece recognition, spatial attention (SA) is introduced into the C2f module of the backbone network of the yolov8 model to achieve the piece recognition. Finally, the location of the Gomoku piece is completed by comparing the position of the piece and the corner point. After experimental validation, the improved model shows a marked improvement over the original yolov8 model, with the precision rate, recall rate, mAP50 and mAP50-95 being 99.9%, 99.0%, 99.4% and 96.4%, respectively. Even under challenging conditions such as uneven brightness, low brightness and tilting of the camera, the recognition and position judgment of the Gomoku pieces can still maintain high accuracy. The introduced method has high performance in Gomoku piece recognition and localisation, providing favourable support for visual detection of Gomoku gaming devices.

Keywords: human-machine game; image processing; yolov8; attention mechanism; piece recognition; localisation.

DOI: 10.1504/IJICT.2025.145830

International Journal of Information and Communication Technology, 2025 Vol.26 No.9, pp.43 - 63

Received: 29 Sep 2024
Accepted: 06 Dec 2024
Published online: 28 Apr 2025 *

Title: Gomoku game recognition and localisation using image processing and deep learning

Keep up-to-date