Realtime Skin Problems Detection with YOLOv5

This project implements real-time skin problem detection using YOLOv5 and a Flask-based web application. The model can detect skin conditions from images, videos, or live webcam feeds.

Features

• Real-time detection: Supports image and video input, including live webcam feeds.
• Flask-based web app: Simple web interface to upload images or videos for analysis.
• YOLOv5-based detection: Custom-trained YOLOv5 model for skin problem classification.

Project Structure

Video_Frame_Split/  
│── model/                   # Pretrained YOLOv5 models (best.pt, best8.pt, etc.)  
│── static/                  # Static files (cover image, PyTorch logo, styles)  
|── templates/index.html      # Frontend (HTML template for Flask app)  
|── main.py                   # Main Flask application script  
│── yolov5s.pt                # YOLOv5 model (this is not the current model used in main.py)  
│── README.md                 # Documentation  

Installation

1. Clone the repository

   git clone https://github.com/ReinerJasin/Realtime-Skin-Problems-Detection-with-YOLO-V5.git
   cd Realtime-Skin-Problems-Detection-with-YOLO-V5

2. Install dependencies (Create a requirements.txt if missing)

   pip install torch torchvision torchaudio flask opencv-python pillow

3. Download YOLOv5 repository

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

Usage

Run the Flask App

python main.py

Then, open http://127.0.0.1:5000/ in your browser.

Upload Images or Videos

• Click on the upload button to submit an image or video for processing.
• The model processes and displays results directly in the browser.

Webcam Detection

• Visit http://127.0.0.1:5000/video_feed to see real-time predictions.

Model Details

• The model is loaded using:

  model = torch.hub.load('ultralytics/yolov5', 'custom', path="model/best.pt", force_reload=True)

• best.pt is the custom-trained YOLOv5 model for skin problem detection.

License

This project is licensed under the Apache License 2.0.

You are free to use, modify, and distribute this project, but attribution is required. Please provide proper credit by linking to this repository when using the project or its components.

For more details, see the LICENSE file.

---