HUE Vision: Eye Tracking using TensorFlow JS and clmtrackr

HUE Vision is a web-based application that harnesses the power of eye tracking technology using web technologies and machine learning. This project serves as a demonstration of real-time gaze prediction directly in the browser.

Demo

Check it out: https://simplysuvi.github.io/hue-vision/

Features

• Real-time eye tracking using the webcam
• Integration of machine learning with clmtrackr.js for facial feature detection
• Training and prediction of gaze direction using TensorFlow.js
• Heatmap visualization for model evaluation

Technologies Used

• JavaScript
• TensorFlow.js
• clmtrackr.js
• HTML/CSS
• jQuery

Getting Started

1. Clone this repository to your local system.
2. Open index.html in a web browser that supports getUserMedia (e.g., Chrome or Firefox).
3. Follow the on-screen instructions to start the webcam and begin eye tracking.

How to Use

1. Allow access to your webcam when prompted.
2. Position your face within the webcam frame.
3. Follow the instructions to start training the machine learning model.
4. Use your gaze to interact with the screen in real time.
5. Explore the heatmap visualization to evaluate the model's performance.

Credits

• This project was inspired by advanced eye tracking technologies such as Apple's Vision Pro.
• Special thanks to the creators of clmtrackr.js and TensorFlow.js for their invaluable contributions to computer vision and machine learning in the browser.