Skip to content

FitDiT

High-fidelity virtual try-on with garment detail

Visit official site

What FitDiT is

The method is described as a high-fidelity virtual try-on approach using a garment texture extractor, frequency-domain learning and a dilated relaxed mask strategy.

What you can do with it

  • Preview a design on a model before production
  • Let shoppers try on clothes before buying
  • Make a virtual try-on video for styling

What to watch out for

  • A clothing image and a person image are both uploaded
  • Parameters such as clothing category are set first
  • A 1024x768 image takes 4.57 seconds to infer

Pros & cons

✓ What we like

  • Stripes, patterns and text details are captured
  • Cross-category try-on is handled by the mask strategy

! What to watch out for

  • A clothing image and a person image are both uploaded
  • Parameters such as clothing category are set first

FAQ

What problem does it address?

Fidelity and robustness in image based virtual try-on.

What inference time is reported?

4.57 seconds for a 1024x768 image.

Last reviewed: 2026-11-15

More AI image generation tools

View all →

How we review