HAND GESTURE CONTROLLED PRESENTATION SYSTEM

Abstract

This project presents a real-time Hand Gesture Controlled Presentation System developed using OpenCV and Python, designed to offer a touch-free and intuitive alternative to conventional presentation control devices such as keyboards, remotes, and clickers. The system leverages MediaPipe’s efficient hand landmark detection framework to accurately track hand movements and extract key landmark points in real time. These landmarks are processed using gesture classification logic to interpret user intentions and translate them into presentation control commands. The proposed system enables seamless slide navigation through natural hand gestures, including moving to the next or previous slide and activating a virtual pointer, thereby improving presenter mobility and audience engagement. Emphasis is placed on system robustness, low latency, and high recognition accuracy across varying lighting conditions, hand orientations, and user differences. To enhance reliability, a custom gesture dataset was collected, and temporal smoothing and debounce mechanisms were implemented to reduce false detections and unintended command execution. Experimental evaluation demonstrates that the system performs effectively in real-world presentation scenarios, maintaining consistent responsiveness and accuracy without the need for specialized hardware. The solution is cost-effective, easy to deploy, and compatible with standard webcams, making it suitable for educational       institutions,     corporate environments,         remote presentations,  and accessibility-focused applications. Overall, this work highlights the potential of computer vision–based gesture control as a practical and natural human–computer interaction technique for modern presentation systems.