AuraFit: Clinical Skin AI & Chromatic Lab

💡 Inspiration

Understanding skin barrier health and finding colors that complement your natural undertones is often overwhelming for everyday consumers. Most people guess their skin type, leading to:

1. Purchasing harsh or mismatched skincare products that cause barrier irritation.
2. Wearing clothing and makeup colors that clash with or wash out their complexion.

I built AuraFit Skin AI to eliminate guesswork by bridging clinical dermatological computer vision with generative AI consultation.

🔬 What It Does

AuraFit delivers an end-to-end, science-backed skin health and chromatic analysis platform in under two seconds:

• 15 Clinical Dermal Biomarkers: Evaluates Moisture/Hydration, Vascular Redness (Erythema), Sebum/Oiliness, Radiance, Acne, Wrinkles, Dark Circles, Pores, Firmness, Perceived Skin Age, and Fitzpatrick Phototype.
• Interactive Facial Zone Heatmapping: Targeted region-by-region diagnostic breakdown across the Forehead, T-Zone, Cheeks, Under-Eye, and Chin.
• Biometric Dermal Radar: Multi-axis visual representation mapping overall barrier vitality and clinical metrics.
• Chromatic Undertone & Seasonal Color Analysis: Accurately classifies undertone (Neutral, Cool, Warm, Olive) and maps optimal seasonal palettes (Deep Winter, Cool Summer, Warm Autumn, Bright Spring) with flattering and unflattering color swatches.
• Dermatologist-Grade AM/PM Regimen: Formulates custom morning antioxidant defense, evening barrier restoration protocols, and flagged harmful ingredients to avoid.
• Multimodal AI Dermo-Aesthetician Agent: A context-aware conversational consultant powered by Google Gemini to answer routine, ingredient, and lifestyle questions.
• Live YouCam API Gateway Inspector: Built-in interactive console for judges to trigger live YouCam REST endpoints and inspect real-time latency and raw JSON payloads.

⚙️ How I Built It

• Core AI & Computer Vision:
  • Perfect Corp YouCam S2S API: Integrated the /s2s/v2.0/task/skin-analysis endpoint for clinical dermal parameter extraction.
  • Google Gemini AI: Implemented context-aware prompt engineering with dynamic model routing (gemini-2.5-flash and gemini-3.7-flash) for real-time aesthetic and ingredient consultation.
• Frontend Architecture:
  • React 19, TypeScript, Tailwind CSS v4, Lucide Icons, and Motion for smooth, high-contrast visual diagnostics.
• Backend Architecture:
  • Node.js & Express API Gateway proxying server-to-server requests to ensure API credentials remain secure.

🧗 Challenges I Ran Into

• Multi-Zone Dermal Mapping: Synchronizing 15 distinct facial biomarkers with dynamic SVG zone overlays across diverse skin tones and facial structures.
• Context Synchronization: Ensuring the conversational AI Dermo-Aesthetician seamlessly ingests the biometric payload to provide medically sound, personalized skincare routines without hallucination.

🏆 Accomplishments That I'm Proud Of

• Ultra-fast end-to-end analysis delivering 15 clinical parameters in under two seconds.
• A clean, clinical-grade UI that makes complex dermatological and color science data intuitive and actionable for anyone.
• Direct, fully testable integration with Perfect Corp's YouCam REST APIs.

🔮 What's Next for AuraFit Skin AI

• Longitudinal Barrier Tracking: Logging skin health improvements over 30/60/90 days to verify product efficacy.
• Environmental & UV Index Integration: Dynamically adjusting AM/PM routines based on real-time local humidity, pollution, and UV exposure.
• Cosmetic Formulation Matchmaker: Connecting diagnosed skin needs directly to real-time e-commerce product catalogs with allergen safety checks.

Built With

• ai-stylist
• apparel-vto
• artificial-intelligence
• cloud-run
• computer-vision
• dermatology
• express.js
• fashion-tech
• gemini-api
• node.js
• perfect-corp
• react
• rest-api
• skin-ai
• tailwindcss
• typescript
• virtual-try-on
• vite
• youcam-api