Virtual Try-On Garment Masking Pipeline with SAM and YOLOS-Fashionpedia

Upload a person image and a garment image to generate a mask for a virtual try-on application. The system will:

  1. Detect the person using YOLO
  2. Create a high-quality segmentation using SAM (Segment Anything Model)
  3. Classify the garment type using YOLOS-Fashionpedia
  4. Generate a mask of the area where the garment should be placed

Note: This system uses state-of-the-art AI segmentation and fashion detection models for accurate results.
Mask Color
0.1 0.9

How It Works

  1. Person Detection: Uses YOLO to detect and locate the person in the image
  2. Segmentation: Uses SAM (Segment Anything Model) to create a high-quality segmentation mask
  3. Garment Classification: Uses YOLOS-Fashionpedia to identify the garment type with fashion-specific detection
  4. Mask Generation: Creates a mask based on the garment type and body part mapping

Supported Garment Types

  • Shirts, Blouses, Tops, and T-shirts
  • Sweaters and Cardigans
  • Dresses and Jumpsuits
  • Skirts
  • Pants, Jeans, and Leggings
  • Shorts
  • Jackets and Coats