Skip to content

README.md

Latest commit

 

History

History
164 lines (118 loc) · 4.5 KB

README.md

File metadata and controls

164 lines (118 loc) · 4.5 KB

👔 FitCheck.AI

Your closet-aware fashion critic.
FitCheckAI is a personal AI stylist that tags your clothes, roasts your fits, and builds outfits - all through an interactive Streamlit UI.

🎥 Demo Video: https://youtu.be/s57yTOkLLwY

🚀 Features

🧠 1. Outfit Analyzer (LangChain + Qwen)

Upload a photo of your outfit and get a brutally honest AI critique, including:

  • Style breakdown
  • Score out of 100
  • Savage one-liner

All results are saved to MongoDB and de-duplicated using perceptual hashing.

🧺 2. Smart Closet Tagging (CLIP)

Upload individual clothing items - the AI classifies:

  • Item type (shirt, shoes, etc.)
  • Color (black, white, etc.)
  • Setting (indoor/outdoor)
  • Formality
  • Gender style

Saved in a consistent JSON format and stored in /Closet.

🧩 3. Outfit Generator

Generates full outfit recommendations based on:

  • Location (indoor/outdoor)
  • Formality (casual/formal)
  • Gender
  • Preferred color

Color compatibility rules are applied (e.g. blue pairs with beige, white, brown, etc).

🛠️ Tech Stack

Component Tool / Library
UI Streamlit
Fashion Critique LangChain + Qwen2.5 VL (3B, quantized)
Image Tagging CLIP (openai/clip-vit-base-patch32)
Database MongoDB Atlas
Data Format JSON / .jsonl

🗂️ Project Structure

.
├── fitcheck/
│   ├── Fashion AI Advisor.py         # Main Streamlit app
│   ├── analyze_outfit.py             # Outfit critic logic
│   ├── tagging.py                    # Tagging with CLIP
│   ├── tag_closet_items.py           # Bulk closet tagger
│   ├── test_Analyze_Outfit.py        # Unit test for LangChain outfit critique
│   ├── test_tag.py                   # Unit test for tagging module
│   ├── testmongoconnection.py        # MongoDB connection test
│   └── pages/
│       ├── 1_Add_to_Inventory.py     # Upload and tag closet items
│       └── 2_Get_Outfit_Suggestion.py# Outfit recommender (rule-based)
├── Closet/                           # JSON-tagged clothing items
├── Images/                           # Outfit photos
├── designs/                          # Wireframes + mockups
├── requirements.txt
├── run_app.bat
└── vlm_tagging_test.ipynb            # Prototype testing

🧪 Example Output

Critique Output

Outfit Critique

Style: The outfit features a casual yet trendy ensemble consisting of a black beanie hat, a white scarf wrapped around the neck, a striped long-sleeve shirt under a black vest over jeans. The combination suggests an urban, laid-back vibe but lacks depth due to its simplicity.

Rating: 35/100

Comment: "This outfit might as well have been designed by a robot; it's so formulaic."

Closet JSON Schema

{
  "image_id": "abc123",
  "item_type": "Jacket",
  "color": "Black",
  "indoor_outdoor": "Outdoor",
  "formality": "Casual",
  "gender": "Men's",
  "path": "Closet/jacket.jpg",
  "folder": "Closet"
}

Outfit Recommender

Outfit Recommender

🖼️ UI Concept Designs

These were mockups used to guide layout and flow.


Landing Page


Upload Interface


LeCritique Display

⚙️ Usage

Install dependencies:

pip install -r requirements.txt

Then launch the app:

streamlit run "fitcheck/Fashion AI Advisor.py"

👥 Team

  • Daniel Mehta
  • George Fotabong
  • Dylan Higuchi
  • Jacob DeAbreu

🧩 Notes

  • MongoDB URI is hardcoded for now (demo only - no sensitive data).

📄 License

This code is free to use, modify, and share for non-commercial purposes.
Commercial use requires prior written permission. All rights reserved.