FashionAI - Clothing Recognition with Color Classification

Key Features & Technical Highlights:

1. Dual-Stage AI Pipeline: Leverages a sophisticated two-stage inference process. First, it identifies clothing bounding boxes via YOLOv8, then classifies the specific color of the cropped item to produce combined labels like "White T-Shirt" or "Blue Jacket".
2. Standardized Labeling: Implements a professional class-mapping system to ensure consistent, user-friendly labels (e.g., displaying "T-Shirt" and "Shorts" instead of raw model outputs like "tshirt" or "short").
3. 8 Clothing Categories & 8 Colors: Trained to accurately recognize:
   • Categories: T-Shirt, Dress, Jacket, Pants, Shirt, Shorts, Skirt, and Sweater.
   • Colors: Beige, Black, Blue, Gray, Green, Pattern, Red, and White.
4. Dual Interaction Modes: Offers flexibility with a Single-Image Capture Mode for precise analysis and a Continuous Live Detection Mode for real-time streaming applications.
5. Premium Web Interface: Features a modern dark-themed UI with glassmorphism effects, smooth animations, drag-and-drop upload functionality, and a responsive design optimized for both desktop and mobile.
6. Modular Clean Architecture: Built with a modular backend structure using FastAPI, ensuring maintainability. Includes automatic API documentation (Swagger/ReDoc), health monitors, and batch processing endpoints.
7. Adjustable Confidence Threshold: Users can fine-tune detection sensitivity in real-time via an intuitive slider, allowing for the perfect balance between precision and recall.