DermaBot : Skin Disease Detection & FAQ Chatbot

DermaBot is a web-based AI application that detects skin diseases from user-uploaded images using a CNN classifier. It also includes a FAQ chatbot for common skin health questions. Built as a thesis project combining computer vision and conversational AI.

People often ignore skin changes because they can't assess severity. Medical consultations require time and money, leading to delayed treatment for conditions that could be caught early.

Build an accessible first-pass screening tool that gives users information about visible skin conditions and directs them toward appropriate care.

• CNN-based image classifier trained on skin disease dataset.
• Web interface for image upload and prediction.
• FAQ chatbot for common skin health questions.
• Prediction result display with confidence and condition info.

• Image upload and classification.
• Multi-class skin disease prediction with confidence scores.
• FAQ chatbot for skin health education.
• Simple web interface.

• Convolutional neural network for image classification.
• FAQ retrieval using text matching for common questions.

1. 1User uploads a skin image through the web interface.
2. 2Image is preprocessed and passed to the CNN model.
3. 3Model returns class probabilities for known skin conditions.
4. 4Top prediction is displayed with confidence and basic information.
5. 5User can ask FAQ questions for additional guidance.

• Limited training data for rare skin conditions.
• Ensuring the app clearly communicates it's not a medical diagnosis.
• Handling image quality variation from phone cameras.

• Frame the tool as educational screening, not medical diagnosis.
• Include disclaimers prominently in the UI.
• Keep the model focused on common conditions where data is available.

Working web application with skin disease classification and FAQ chatbot. Thesis project with documented methodology and results.

• Expand training dataset with more diverse skin tones.
• Add specialist referral recommendations based on condition severity.
• Build mobile app version for easier image capture.