StyleSense — FashionAI 👗

StyleSense is an AI-powered personalized fashion recommendation system. Upload your photos and it analyzes your skin tone, eye color, hair color, and body shape, then recommends outfits, color palettes, fabrics, and styling advice — backed by an LLM-powered stylist chatbot.

The project has two components:

1. code2_modified.py — PersonalAnalyzer (computer vision + ML)

   • Uses MediaPipe FaceMesh (with refined iris landmarks) to detect facial regions
   • Extracts dominant skin tone, under-tone, eye color, and hair color from images via clustering
   • Classifies body shape / proportions
   • An XGBoost multi-output classifier maps extracted features to personalized recommendations (clothing colors, fits, materials, patterns, jewelry metals, shoes) using the curated knowledge table in _recommendations_with_hex.csv

2. chatbot_ui.py — Stylist Chatbot (Streamlit + Groq LLM)

   • Polished Streamlit UI for uploading photos and viewing your personalized palette
   • An LLM chatbot (Groq API) that answers styling questions informed by your analysis results

pip install -r requirements.txt

The app reads the key from the GROQ_API_KEY environment variable (get one at console.groq.com):

export GROQ_API_KEY="your-key-here"   # Windows: set GROQ_API_KEY=your-key-here

streamlit run chatbot_ui.py

Upload a front-facing photo, let the analyzer extract your features, and start chatting with your AI stylist.

• Personalized clothing color palette (with hex codes) and colors to avoid
• Recommended fits, fabrics, and patterns for your body shape
• Jewelry metal and shoe style suggestions
• A conversational stylist chatbot for follow-up fashion advice

Python 3.9–3.11. Key dependencies: mediapipe, opencv-python, scikit-learn, xgboost, streamlit, groq, Pillow — see requirements.txt for the full pinned list.