GlamSync
What if your laptop camera could do more than just show you your reflection?
We built GlamSync AI, a personal beauty styling coach that combines Gemma 4, OpenCV, and real-time AR to help users understand what makeup and styling choices actually suit them.
The Problem
Choosing makeup that works with your individual face, skin undertone, and outfit can be surprisingly difficult.
Most beauty applications either provide static filters or generic recommendations. They may tell you what to use, but they don't understand your facial features or show you how to apply it.
We wanted to bridge that gap.
Our Solution
GlamSync AI creates a personalized styling pipeline:
Outfit → Facial Analysis → AI Recommendation → AR Application Guidance
1. Outfit Color Analysis
We use OpenCV and K-Means clustering to extract dominant colors from an outfit and understand its palette.
2. Facial Analysis
OpenCV YuNet analyzes facial geometry while CIELAB-based ITA colorimetry and skin segmentation help estimate skin tone and undertone.
3. Gemma 4 Reasoning
The user's facial profile, outfit colors, occasion, budget, and available products are passed to Gemma 4, which generates personalized styling looks and step-by-step recommendations.
4. Real-Time AR Coach
This is where GlamSync goes beyond a normal recommendation system.
Using MediaPipe's 468-point Face Mesh, our AR mirror tracks facial landmarks and overlays guidance for areas such as the cheeks, eyelids, eyeliner, and lips.
The user can follow the guidance while applying their makeup, with voice instructions for a hands-free experience.
Why Open-Source AI Matters
We wanted AI to be an actual part of the product rather than simply adding a chatbot.
Our pipeline combines open/open-weight AI and local computer vision components so that facial analysis and personalization are tightly integrated into the application.
This also gives us the flexibility to experiment with different models and reduce dependence on a single closed AI provider.
Tech Stack
- React + TypeScript + Vite
- FastAPI + Python
- Gemma 4 Multimodal
- OpenCV + YuNet
- MediaPipe Face Mesh
- NumPy
- CIELAB / ITA color analysis
- Google AI Studio
- WebRTC / Web Speech API
What We Built
The final system takes the user from:
"What suits me?"
to
"What should I use?"
to
"Show me exactly how to apply it."
That's what we wanted GlamSync AI to be — not just an AI beauty recommender, but an interactive AI styling coach.
Built for Hacktoberfest Hack Day Coimbatore
Built by Team GlamSync AI for Hacktoberfest Hack Day Coimbatore.
🔗 GitHub: https://github.com/guruc1234a-spec/hacktoberfest-hack-day-coimbatore-x-init-club-and-idea-club
Top comments (0)