Posture_detection_using_PoseNet_23B2476
This project uses ml5.js and PoseNet to perform real-time human pose detection via webcam and apply fun visual filters like glasses and a cigar based on facial keypoints.
- 📸 Accesses the user's webcam and detects body keypoints using PoseNet
- 🧠 Uses
ml5.jsfor simplified machine learning integration in the browser - 🕶️ Dynamically overlays filters (like glasses or cigar) onto detected facial points
- 🔄 Flipped video rendering to simulate a mirror-like experience
- ✅ Option to toggle filters directly from the UI
- HTML5 / CSS3
- JavaScript
- p5.js — creative coding library for rendering
- ml5.js — friendly ML library built on TensorFlow.js
PoseNet-Filter-Demo/ ├── index.html ├── sketch.js ├── images/ │ ├── spects.png │ ├── cigar.png
- Clone the repository:
git clone https://github.com/yourusername/posenet-filter-demo.git cd posenet-filter-demo - Start a local live server (VSCode extension or Python HTTP server):
python -m http.server
-
Open your browser and navigate to: http://localhost:8000
-
Allow camera access to begin real-time pose tracking.
🧪 Features Real-time pose detection via webcam
Draws keypoints and skeleton overlay
Applies virtual filters using facial landmark positioning
Filter toggle control via checkbox (no page reload)