Vision-Assist 🎯

This project demonstrates real-time object detection using the YOLOv5 model and OpenCV via a webcam feed. It identifies multiple objects in live video streams and draws bounding boxes with class labels and confidence scores.

Overview

• Captures live video from webcam
• Uses a pre-trained YOLOv5 model (yolov5s.pt) for detection
• Detects 80 object classes from the COCO dataset
• Draws bounding boxes with class names and confidence
• TTS is used to describe objects
• Thermal depth verification
• Built with Python, OpenCV, and PyTorch

Technologies Used

• Python
• OpenCV
• PyTorch
• YOLOv5 (via Ultralytics)
• Jupyter Notebook

How to Run Locally

1. Clone the repo

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

2. Install dependencies

   pip install -r requirements.txt

3. Download YOLOv5 repository

   git clone https://github.com/ultralytics/yolov5.git
   cd yolov5
   pip install -r requirements.txt

4. Place your notebook inside the yolov5 folder or adjust the paths accordingly.

5. Run the notebook Open Live_Object_Detection.ipynb and run all cells.

📝 Notes

• Ensure your webcam is accessible and enabled.
• For better performance, consider switching to a GPU runtime (e.g., using Google Colab).
• You can replace yolov5s.pt with other YOLOv5 model weights for higher accuracy (e.g., yolov5m.pt, yolov5l.pt).

📄 License

This project is open-source under the MIT License.