Gesturize: Democratizing Real-Time Hand Gesture Recognition for Accessible Human-Computer Interaction

Vamsi Naghichetty Kishore Kumar, Prithvi Seshadri, Gokul Srinath Seetha Ram

Keywords: Hand Gesture Recognition, Accessibility, Economic Accessibility, Real-time Human-Computer Interaction, MediaPipe, Assistive Technology, Open Source, Multi-modal Interaction, Computer Vision, TensorFlow Lite, Democratization of Technology, Inclusive Design, Low-cost Solutions

TL;DR: Gesturize democratizes gesture recognition by enabling accessible, real-time hand control of computers using only standard webcams, eliminating economic barriers while maintaining technical excellence.

Abstract: We present Gesturize, a real-time hand gesture recognition system designed to democratize accessible human-computer interaction through cost-effective computer vision techniques. While existing gesture control systems require expensive proprietary hardware (e.g., smart glasses costing \$299-\$379), our approach leverages standard webcams and smartphones, eliminating economic barriers to assistive technology adoption. Our system integrates MediaPipe for robust 21-keypoint hand tracking with a custom TensorFlow Lite classifier, achieving 94\% accuracy across nine gesture classes with sub-100ms latency. The multi-modal framework combines neural gesture classification with speech recognition, enabling touchless control particularly beneficial for mobility-impaired users. Through comprehensive evaluation including user studies with individuals across diverse economic backgrounds and accessibility needs, we demonstrate that Gesturize provides comparable functionality to commercial solutions at near-zero hardware cost. Our open-source implementation addresses critical gaps in accessible technology, making advanced gesture recognition available to underserved communities and individuals with disabilities who cannot afford existing solutions.

Submission Number: 12