📡 Radar + Object Detection with TensorFlow.js

This project combines a radar-style canvas visualization with real-time object detection using TensorFlow.js and the COCO-SSD model. It works directly in the browser using the device camera.

🔍 Features

• 🎯 Real-time object detection using COCO-SSD
• 🧠 Option to filter for people only
• 🗣️ Voice output using Speech Synthesis API
• 🎛️ Control panel to pause detection, mute/unmute voice, and toggle filters
• 📡 Radar canvas that simulates motion detection with red blips
• 📷 Hidden camera feed for a clean UI
• ✅ Fully offline support possible with local model files

📦 Installation

Option 1: Run Online

Just open index.html in your browser with internet access (uses CDN for models).

Option 2: Offline Setup

1. Download the COCO-SSD model and save it in a models/coco-ssd/ folder:

   • model.json
   • Binary weight .bin file (linked from above .json)

2. Place both tf.min.js and coco-ssd.min.js in your project folder.

3. Update the script tags in index.html:

   <script src="tf.min.js"></script>
   <script src="coco-ssd.min.js"></script>
     

4. Update main.js to load the model locally : model = await cocoSsd.load({ modelUrl: 'models/coco-ssd/model.json'});

project structure for offline project

├── index.html
├── main.js
├── style.css
├── tf.min.js                # (optional - offline only)
├── coco-ssd.min.js          # (optional - offline only)
└── models/
 └── coco-ssd/
     ├── model.json
     └── group1-shard1of1.bin