Skip to content

Latest commit

 

History

History
49 lines (38 loc) · 1.79 KB

README.md

File metadata and controls

49 lines (38 loc) · 1.79 KB

Hand Gesture Controlled Dino Game using React, TensorFlow.js & Electron

Overview

This project is a hand gesture recognition system that enables users to play the Chrome Dino Game using a specific hand gesture. It leverages computer vision and machine learning to detect hand landmarks and recognize predefined gestures in real time. The application is built using React, TensorFlow.js, and Electron, making it suitable for both web and desktop environments.

Tech Stack

  • React
  • TensorFlow.js
  • HandPose Model
  • Fingerpose
  • Electron
  • Webcam API

Features

✅ Real-time hand tracking using TensorFlow.js HandPose Model
✅ Gesture recognition with Fingerpose
✅ Electron integration for desktop interactions
✅ Play Chrome Dino Game by performing a specific hand gesture
✅ Canvas rendering of hand skeleton for better visualization

Gesture: "Point Up" for Jump

  • The index finger must be fully extended and pointing up.
  • All other fingers must be curled.
  • The gesture estimator confirms the match with a high confidence score before triggering the space key event.

Use Cases

🎮 Gaming - Hands-free control to play the Chrome Dino Game
💻 Accessibility - Assistive technology for users with limited mobility
🖥️ Fun Experiment - Explore AI-powered gesture control

Installation & Setup

# Clone the repository
git clone https://github.com/yourusername/hand-gesture-dino-game.git
cd hand-gesture-dino-game

# Install dependencies
npm install

# Run the application
npm start

Future Improvements

🚀 Add more gestures for different actions
🚀 Implement gesture customization
🚀 Improve detection accuracy with additional ML techniques