AI-powered face tracking experience that generates realistic gaze-following effects. Uses Replicate's Expression Editor model to create 121 face images in an 11×11 grid, then maps cursor/touch position to the appropriate image for real-time interaction.
Image Generation
How Images Are Created:
• Upload a 512×512 face photo to FaceLooker Replicate Model
• AI generates 121 face images looking in different directions (11×11 grid)
• Uses Expression Editor model for realistic gaze variations
• Images cover gaze angles from -15° to +15° in 3° increments
• Downloads complete ZIP with images, sprite sheet, and demo code
Grid System: 11×11 images (-15° to +15°, 3° steps = 121 total) Real-time Tracking: Maps cursor position to appropriate face image Interactive Experience: Creates engaging UI elements for portfolios, avatars, and demos Mobile Support: Full touch gesture compatibility for mobile devices
🛠️ Technical Implementation
Our refactored version advances the original vanilla JS implementation with modular ES6 architecture, event-driven design, PWA features, intelligent caching, theme system, and production optimizations while maintaining the creative AI-powered experience.
This project is licensed under the MIT License. You are free to use, modify, distribute, and create derivative works for both personal and commercial purposes.