Vision Project - Real-time Object Recognition

A plug-and-play object recognition system using your MacBook M2 Pro's built-in camera. This project uses OpenCV for camera access and YOLOv8 (You Only Look Once) for real-time object detection.

Features

• 🎥 Real-time object detection using your MacBook's camera
• 🚀 Fast and efficient YOLOv8 model (optimized for Apple Silicon)
• 📊 Displays bounding boxes, labels, and confidence scores
• 💾 Save frames with detected objects
• ⌨️ Interactive controls

Quick Start

1. Install Dependencies

cd visionProject
pip install -r requirements.txt

Note: On first run, YOLOv8 will automatically download the model weights (~6MB for nano model).

2. Run the Application

python object_recognition.py

3. Controls

• 'q' - Quit the application
• 's' - Save current frame as an image
• 'r' - Toggle object detection on/off

How It Works

1. Camera Access: Uses OpenCV to access your MacBook's built-in camera
2. Object Detection: YOLOv8 model processes each frame to detect objects
3. Visualization: Draws bounding boxes and labels around detected objects
4. Real-time Processing: Optimized for smooth real-time performance on M2 Pro

Model Options

The default uses YOLOv8n (nano) for fastest performance. You can change the model size in object_recognition.py:

recognizer = ObjectRecognizer(model_size='n')  # Options: 'n', 's', 'm', 'l', 'x'

• 'n' (nano) - Fastest, ~6MB, good for real-time
• 's' (small) - Balanced, ~22MB
• 'm' (medium) - Better accuracy, ~52MB
• 'l' (large) - High accuracy, ~87MB
• 'x' (xlarge) - Best accuracy, ~136MB

Requirements

• Python 3.8+
• macOS (tested on MacBook M2 Pro)
• Camera permissions (granted automatically on first run)

Troubleshooting

Camera not working?

• Make sure no other application is using the camera
• Grant camera permissions in System Settings > Privacy & Security > Camera

Slow performance?

• Use the nano model ('n') for fastest performance
• Close other applications using the camera
• Reduce camera resolution in the code if needed

Model download issues?

• Ensure you have an internet connection for the first run
• The model will be cached after first download

What Objects Can It Detect?

YOLOv8 can detect 80 different object classes including:

• People, animals (cats, dogs, birds, etc.)
• Vehicles (cars, trucks, buses, motorcycles, etc.)
• Electronics (cell phones, laptops, keyboards, etc.)
• Furniture (chairs, couches, tables, etc.)
• Sports equipment (sports balls, skateboards, etc.)
• And many more!

Example Output

The application will display:

• Live video feed from your camera
• Bounding boxes around detected objects
• Object labels and confidence scores
• Frame counter and status information

License

This project is open source and available for personal and educational use.