Skin-Lab-Rx

Skin Lab Rx uses Perfect Corp AI to provide 14-metric skin diagnostics, tailored product matching, and virtual makeup try-on, turning any selfie into a personalized, clinical skincare experience.

Comment

Inspiration

The skincare industry is overwhelming, with thousands of products and generic advice. I was inspired to build Skin-Lab-Rx to bridge the gap between clinical dermatology and at-home beauty. By leveraging high-fidelity AI, I wanted to give everyone a "digital dermatologist" in their pocket that doesn't just guess, but truly analyzes.

Skin-Lab-Rx: Hackathon Submission Content

What it does

Skin-Lab-Rx is a comprehensive skincare platform that:

  • Analyzes: Scans 14 unique skin metrics (from hydration to eye bags) via a high-precision AI selfie scan.
  • Diagnoses: Provides a "Skin Age" and prioritized concern breakdown.
  • Prescribes: Intelligently maps products to your specific diagnostic results.
  • Visualizes: Uses AI Makeup Transfer for a Virtual Try-On experience to see results before you buy.

How we built it

I built a modern, mobile-first SPA using:

  • Next.js 16 and TypeScript for a robust, performant foundation.
  • Perfect Corp AI APIs for the heavy lifting: AI Skin Analysis (v2.0) and AI Makeup Transfer (v2.0).
  • Vanilla CSS: A custom design system featuring glassmorphism and premium dark-mode aesthetics.
  • Frontend Logic: Custom face-detection and auto-cropping algorithms to ensure high-quality API inputs.

Challenges we ran into

The biggest challenge was the strictness of clinical AI. The Perfect Corp API requires very specific face ratios and lighting. I had to build a custom Smart Alignment system with a draggable UI and a resilient backend fallback to ensure that "real world" selfies still produced beautiful, actionable results without failing or crashing.

Accomplishments that we're proud of

  • The UI/UX: Creating a draggable alignment guide that feels native and intuitive.
  • Resilience: Implementing a backend that gracefully handles API clinical errors so the user never sees a failure.
  • Actionability: Building the "Concern-to-Product" mapping that makes the analysis feel meaningful and useful.

What we learned

I learned the intricacies of working with professional-grade AI APIs—specifically the importance of data pre-processing (cropping, scaling, and alignment) before sending data to an AI model. I also mastered the "FaceDetector" browser API for real-time user guidance.

What's next for Skin-Lab-Rx

I plan to expand the product database, integrate with e-commerce platforms for instant checkout, and add AR-guided application routines that show you exactly where to apply serums and treatments based on your scan's heatmaps.

Share this project:

Updates