MaixCAM-based Tic-Tac-Toe human-computer game system design

In this paper a human-machine Tic-Tac-Toe gaming system is constructed, which is based on the MaixCAM vision processing platform. This game system integrates real-time image recognition and intelligent decision-making. An OS04A10 camera is used as the visual information acquisition module and a robotic arm composed of servo motors and linear actuators driven by PID algorithms is used as the execution device. Different from the traditional minimax algorithm for Tic-Tac-Toe, this system dynamically adapts the weighting algorithm from Gomoku (Five-in-a-Row) to Tic-Tac-Toe, establishing a lightweight decision-making framework based on local threat assessment. The system enables human-machine gameplay with both first and second player turns and supports solving endgame situation.