SkinLens: Interactive Skin Lesion Diagnosis using CNNs

SkinLens is a web application designed to assist in the diagnosis of skin lesions. Users can upload high-quality images of skin lesions, which are then classified using a convolutional neural network. This tool aims to provide an interactive and accurate diagnosis suggestion to help in early detection and treatment.

Project Structure

skin-lens/
│── .gitignore              # Git ignore file
│── LICENSE                 # License file
│── README.md               # Project documentation
│── api/                    # Backend server for model inference
│   ├── main.py             # FastAPI entry point
│   ├── readme.md           # API specific documentation
│   ├── requirements.txt    # Backend dependencies
│   ├── models/             # Model (trained weights, model loading)
│   │   ├── __init__.py
│   │   ├── checkpoint_epoch__effnet_b1_35.pth
│   │   ├── cnn_model.py
│   ├── routes/             # API routes for handling requests
│   │   ├── __init__.py
│   │   ├── classify_routes.py
│   ├── utils/              # Utility functions (preprocessing)
│       ├── __init__.py
│       ├── image_processing.py
│── frontend/               # Frontend application
│   ├── .babelrc            # Babel configuration
│   ├── package.json        # Frontend dependencies
│   ├── webpack.config.js   # Webpack configuration
│   ├── public/             # Static assets
│   ├── src/                # React components
│       ├── App.js          # Main frontend app
│       ├── index.js        # Entry point
│       ├── index.css       # Global styles
│       ├── components/     # UI components
│       ├── pages/          # React pages
│           ├── Home.jsx
│           ├── LandingPage.jsx
│── notebooks/              # Jupyter Notebooks for model training
│   ├── resnet18_transfer_isic2019.ipynb

API Description

The SkinLens API is built using FastAPI and serves a Convolutional Neural Network (CNN) model for image classification. The model used is EfficientNet-B1, which has been trained on the ISIC 2019 dataset for skin lesion classification (multiclass classification, 8 classes). The API provides endpoints for image classification and is designed to handle image uploads, preprocess the images, and return classification results.

How It Works

1. Image Upload: Users upload an image of a skin lesion through the frontend.
2. Image Preprocessing: The uploaded image is preprocessed to match the input requirements of the CNN model.
3. Model Inference: The preprocessed image is passed through the EfficientNet-B1 model to obtain classification results.
4. Result: The classification results, including the predicted class and confidence score, are returned.

API Endpoints

• GET /: Returns a welcome message.
• POST /api/classify: Accepts an image file and returns the classification result (class and confidence score).

How to Run Locally

API

1. Create a Virtual Environment

   First, create a virtual environment to isolate the project's dependencies. You can use venv for this purpose.

   On Unix-based systems:

   python3 -m venv venv
   source venv/bin/activate

   On Windows:

   python -m venv venv
   venv\Scripts\activate

2. Install Dependencies

   With the virtual environment activated, install the required dependencies using the requirements.txt file.

   pip install -r api/requirements.txt

3. Run the FastAPI Application

   Use uvicorn to run the FastAPI application.

   uvicorn api.main:app --reload

   The server will start, and you can access the API at http://127.0.0.1:8000/api/classify.

Frontend

1. Navigate to the Frontend Directory

   cd frontend

2. Install Project Dependencies

   npm install

3. Run the Webpack Server in Development Mode

   npm run dev

   The frontend application will start, and you can access it at http://localhost:3000.

Additional Information

Frontend Components

• App.js: Main application component that sets up routing.
• Home.jsx: Page component for the skin lesion classifier.
• LandingPage.jsx: Landing page component.
• Button.js: UI component for buttons.
• Card.js: UI component for cards.
• CardContent.js: UI component for card content.

Styling

• Tailwind CSS: Used for utility-first CSS styling.
• Custom Styles: Defined in src/index.css.

Development Tools

• Webpack: Module bundler for the frontend.
• Babel: JavaScript compiler for using the latest JavaScript features.
• PostCSS: Tool for transforming CSS with JavaScript plugins.

License

This project is licensed under the MIT License - see the LICENSE file for details.