Skripsi
PENGEMBANGAN GAME MULTIPLAYER GUNTING BATU KERTAS BERBASIS DETEKSI GERAKAN TANGAN MENGGUNAKAN MEDIAPIPE DAN KOMUNIKASI REALTIME DENGAN SOCKET.IO
In the era of computer vision and machine learning advancement, the integration of gesture recognition technology in interactive gaming continues to evolve. This study aims to develop a multiplayer rock-paper-scissors game based on hand gesture detection using MediaPipe and real-time communication with Socket.IO. The MediaPipe Hands model is utilized to detect and classify hand gestures, while Socket.IO facilitates real-time data exchange between players. Testing was conducted involving 10 different players across 10 game sessions with a total of 50 gameplay rounds on a local area network. Evaluation results demonstrate that the gesture detection system achieves 100% accuracy for all gesture types (rock, paper, scissors) with an average Confidence score of 0.9276 and an average processing time of 68.85 ms. The local multiplayer communication system exhibits excellent performance with an average latency of 18.44 ms, 100% synchronization rate, and 0% disconnect rate throughout all testing sessions. The integration of both systems produces a total response time of approximately 87 ms from gesture detection to game result display. The developed system effectively provides a responsive and stable multiplayer gaming experience, successfully meeting all research objectives and exceeding established success criteria. Key Words: MediaPipe, Socket.IO, Gesture Recognition, Multiplayer Game, Real-Time Communication.
No other version available