Melanoma Detection App

Overview

This Melanoma Detection App is an interactive web application designed to assist in the detection and understanding of melanoma, a serious form of skin cancer. The app utilizes advanced deep learning models to analyze skin and dermoscopy images, providing educational resources and visualizations to enhance understanding of melanoma.

Features

• Multiple Model Support: Includes various deep learning models (CNN, VGG16, ResNet50, EfficientNetB4, InceptionResNetV2) for both skin and dermoscopy image analysis.
• Interactive Image Upload: Users can upload their own skin or dermoscopy images for analysis.
• Model Performance Metrics: Displays detailed performance metrics including accuracy, precision, recall, and AUC for each model.
• Visualizations: Offers interactive data visualizations to understand the distribution of melanoma cases, age, and gender in the dataset.
• Educational Resources: Provides curated links to academic articles, websites, and research papers on melanoma and related technologies.
• FAQs Section: Answers common questions about the app and melanoma detection.
• Feedback System: Includes a form for users to submit feedback or questions.

Technology Stack

• Frontend: Streamlit
• Backend: Python
• Deep Learning: TensorFlow, Keras
• Data Visualization: Plotly, Seaborn
• Image Processing: Pillow

Installation

1. Clone the repository:

   git clone https://github.com/DeborahAdedigba/Melanoma-Detection.git

2. Install the required dependencies:

   pip install -r requirements.txt

3. Run the Streamlit app:

   streamlit run app.py

Usage

Navigate through the app using the sidebar menu:

1. Introduction: Learn about the app and melanoma.
2. Model Performance: Explore the performance metrics of different models.
3. Visualizations: View interactive charts and graphs of melanoma data.
4. Melanoma Detection: Upload images for analysis.
5. Educational Resources: Access curated learning materials.
6. FAQs: Find answers to common questions.
7. Feedback and Contact: Submit your feedback or questions.

Data Sources

• PH2 Dataset for skin images
• ISIC 2016 Dataset for dermoscopy images

Disclaimer

This app is for educational and informational purposes only. It is not a substitute for professional medical advice, diagnosis, or treatment. Always seek the advice of your physician or other qualified health provider with any questions you may have regarding a medical condition.

Contributing

Contributions to improve the app are welcome. Please feel free to submit pull requests or open issues to discuss potential enhancements.

License

GNU GENERAL PUBLIC LICENSE Version 3, 29 June 2007

Contact

For any queries or support, please use the Feedback and Contact form within the app or reach out to 2adedd38@solent.ac.uk.