Real-Time Hand Gesture Recognition for Controlling Web Video of Traditional Dance Utilizing MediaPipe

Abstract

The interaction between humans and digital devices has significantly evolved with a notable progression in Human-Computer Interaction (HCI) being the increasing adoption of contactless device control through hand gesture recognition. This research details the development and evaluation of a real-time hand gesture recognition system utilizing the MediaPipe framework for controlling web video players, with a specific focus on its application in managing the playback of traditional dance videos. The system enables users to manipulate video playback functions via hand gestures. The research methodology comprised key stages: a literature review, system workflow design, practical implementation, and performance evaluation. The MediaPipe framework facilitated real-time detection of hand landmark coordinates, processed, and mapped to specific video player control commands (e.g., play, pause, volume adjustment) using a rule-based approach. The findings indicate that the system performs very well in bright lighting conditions and at close user distances (0.5 meters) from the camera, accurately recognizing gestures with high responsiveness. However, recognition accuracy significantly decreases at further distances (>3 meters) and in dark lighting conditions, highlighting the importance of optimal camera placement and sufficient illumination for effective system operation.  The evaluation results indicate effective control of web video players, demonstrating optimal performance in environments with bright illumination and close user proximity.