GlamSync
GlamSync is a personal beauty styling coach that combines computer vision, Gemma 4, and real-time AR to help users discover and apply looks tailored to them.

Users upload an outfit and capture a selfie. GlamSync analyzes the outfit’s color palette using K-Means clustering, detects facial morphology and skin undertone using OpenCV YuNet and CIELAB-based ITA analysis, and passes these insights to Gemma 4 multimodal reasoning. Gemma generates three personalized looks based on the user’s face, outfit, occasion, budget, and available products.

The experience goes beyond recommendations. Our real-time AR Mirror Coach uses MediaPipe’s 468-point Face Mesh to track facial landmarks and project step-by-step guides for areas such as cheekbones, eyelids, eyeliner, and lips, accompanied by voice instructions.

GlamSync AI brings the complete journey together:

“What suits me?” → “What should I use?” → “Show me how to apply it.”

By combining local computer vision with Gemma 4, the project turns a standard camera into an interactive, personalized beauty coaching experience.