Hand Gesture based AI Controller for Presentation, Virtual Drawing and System Volume Management

The use of wireless control systems for human-computer interaction has grown in popularity recently. Earlier, people used to stick to input tools like a keyboard or mouse, which restrict one’s movement and interaction options, but now, newer technologies like computer vision, machine learning and AI have opened new and interactive ways of interacting with computers. This paper presents a hand gestures approach that not only can control presentations but is also useful for facilitating drawing and managing system volume which enrich the user experience and engagement by offering an interface that eliminates the need for keyboard and mouse. With the help of this algorithm, the computer system accurately responds to hand movements in real time. With this model, one can easily navigate through slides, play, or pause videos, and perform multiple presentation related tasks using hand gestures. It boosts mobility and enhances audience engagement by removing the limitations posed by traditional input methods. It also allows users to digitally draw on the system. Users can sketch diagrams, annotate slides, and unleash their creativity without requiring drawing tablets. It uses Google’s mediapipe module to detect hand landmarks. We incorporated specific gestures that enable the model to initiate a particular action. It showed an accuracy of $\mathbf{9 8 \%}$ under normal lighting conditions, which slightly drops to $95 \%$ and $\mathbf{9 6 \%}$ in low and bright light, respectively, while the overall accuracy remains at $96.33 \%$.