Inspiration

Skin inflammations are common, yet many users struggle to understand visible skin changes and have limited access to dermatologists. This project was inspired by the idea that short video analysis combined with multimodal AI can help users better observe their skin without providing medical diagnosis.

What it does

Smart Skin Insight analyzes short skin videos (5–10 seconds) to observe visual changes over time, including:

  • redness
  • texture variations
  • swelling

Leveraging the power of Gemini 3, the app performs multimodal video analysis and temporal reasoning to identify patterns and generate:

  • non-medical categorization
  • clear explanations
  • adaptive skin care routines
  • lifestyle and nutrition guidance

How we built it

  • backend: Django REST API for video processing
  • frontend: React web application with CSS and Tailwind for styling
  • AI: Gemini 3 for multimodal video analysis and temporal reasoning

Challenges we ran into

  • analyzing temporal visual changes instead of static images
  • ensuring all outputs remain strictly non-medical
  • making AI insights understandable for non-expert users

Accomplishments that we're proud of

  • video-based multimodal analysis
  • temporal reasoning on short clips
  • complete end-to-end web application

What we learned

We learned how to apply multimodal AI responsibly, especially in health-related contexts where clarity, limitations, and ethics are critical.

What's next for Smart Skin Insight

  • tracking skin evolution across multiple sessions
  • improved personalization
  • expert-reviewed, non-medical guidance

Built With

  • css
  • django-(backend)
  • gemini
  • next.js
  • react-(frontend)
  • tailwind
Share this project:

Updates