Skip to main content

Garment Segmentation

Detailed guide to garment segmentation using U2Net.

Overview​

Garment segmentation identifies and separates different garment types from images.

Usage​

from tryon.preprocessing import segment_garment

segment_garment(
    inputs_dir="data/original_cloth",
    outputs_dir="data/garment_segmented",
    cls="upper"
)

Parameters​

  • inputs_dir: Directory containing input images
  • outputs_dir: Directory to save segmentation masks
  • cls: Garment class - "upper", "lower", "dress", or "all"

Output​

Segmentation masks saved as JPG files with format: {image_name}_{class_id}.jpg

See Preprocessing Overview for more details.