How fareed7479/oral-cancer works
OncoDetectAI is an oral-cancer screening web application. A user uploads a clinical image; the frontend requests a prediction, then displays the classification, confidence, and image visualizations. Users can continue to a lifestyle assessment and recommendations flow. The backend excerpt shows image handling, model inference, and Grad-CAM generation; the lifestyle endpoint is shown only as commented-out code, so its connection to Gemini is unverified. The README is empty.
Main entry points: App.jsx Route dispatch · HomePage.jsx Image upload · main.py FastAPI endpoints · ResultPage.jsx Diagnostic report
Components 11
Screening interface
- Image upload
frontend/src/components/HomePage.jsx - Diagnostic report
frontend/src/components/ResultPage.jsx - Route dispatch
frontend/src/App.jsx
Screening backend
- FastAPI endpoints
backend/main.py - Image prediction
backend/main.py - Grad-CAM overlays
backend/main.py
Lifestyle guidance
- Lifestyle assessment
frontend/src/pages/LifestyleAssessment.jsx - Recommendations view
frontend/src/pages/LifestyleResult.jsx
AI integration
- Recommendation generation
backend/services/gemini_service.py
Other components
- User
- Google Gemini
Connections 13
- User → Image upload: uploads image
frontend/src/components/HomePage.jsx - Image upload → FastAPI endpoints: requests prediction
frontend/src/components/HomePage.jsx - FastAPI endpoints → Image prediction: dispatches inference
backend/main.py - Image prediction → Grad-CAM overlays: generates overlays
backend/main.py - Image upload → Diagnostic report: navigates with result
frontend/src/components/HomePage.jsx - Diagnostic report → Lifestyle assessment: opens assessment
frontend/src/components/ResultPage.jsx - Lifestyle assessment → FastAPI endpoints: requests plan
frontend/src/pages/LifestyleAssessment.jsx - FastAPI endpoints → Recommendation generation: may invoke
backend/main.py - Recommendation generation → Google Gemini: generates content
backend/services/gemini_service.py - Route dispatch → Image upload: routes to
frontend/src/App.jsx - Route dispatch → Diagnostic report: routes to
frontend/src/App.jsx - Route dispatch → Lifestyle assessment: routes to
frontend/src/App.jsx - Route dispatch → Recommendations view: routes to
frontend/src/App.jsx