Gesture-Based Air Writing System Utilizing Computer Vision

Abstract

The Gesture-Based Air Writing System Utilizing MediaPipe and Computer Vision presents an innovative solution for air-writing recognition, leveraging artificial intelligence (AI) techniques to provide an efficient and user-friendly experience. This project integrates MediaPipe, a real-time hand tracking library, to accurately capture air-writing gestures from users using a single web camera. The system eliminates the need for physical constraints such as delimiters or imaginary boundaries, allowing unrestricted hand movements. A custom preprocessing pipeline is employed to convert the captured hand trajectory data into suitable formats, which are then processed by deep learning models, particularly convolutional neural networks (CNNs), for precise character recognition. MediaPipe’s hand tracking algorithm enhances the robustness and accuracy of the system by detecting and following finger movements in real time. The use of CNNs ensures that the system can effectively recognize air-written characters while maintaining high accuracy.