React Native Object Detection App

A real-time object detection app built with React Native using VisionCamera and TensorFlow Lite. Detect 80+ objects in real-time with bounding boxes and confidence scores.

✨ Features

• 📱 Real-time object detection with live camera feed
• 🎯 Bounding boxes with confidence percentages
• 📷 Front/back camera switching
• ⚡ FPS monitoring and performance stats
• 🎚️ Adjustable confidence threshold (10-90%)
• 🏷️ 80+ object classes from COCO dataset

📱 Screenshots

Real-time object detection with bounding boxes and confidence scores

🚀 Quick Start

1. Clone the repository

   git clone https://github.com/yourusername/react-native-object-detection.git
   cd react-native-object-detection
   

2. Install dependencies

   npm install
   

3. Run the app

   # iOS
   npx expo run:ios
   
   # Android
   npx expo run:android
   

🛠️ Tech Stack

• React Native 0.79.3 - Cross-platform mobile framework
• Expo - Development platform with prebuild workflow
• VisionCamera - Camera access and frame processing
• TensorFlow Lite - On-device ML inference
• SSD MobileNet v1 - Pre-trained object detection model

🎯 How It Works

1. Camera Feed: Captures live video frames at 30 FPS
2. Preprocessing: Resizes frames to 300x300 pixels
3. Inference: Runs TensorFlow Lite model on each frame
4. Postprocessing: Filters detections by confidence threshold
5. Rendering: Draws bounding boxes with labels on screen

⚙️ Configuration

Adjusting Performance

• Confidence Threshold: Use the slider in the app (10-90%)
• Frame Processing: Modify frame processor in App.js
• Model Quality: Replace with different TensorFlow Lite models

Custom Models