Virtual Try-On V2: 6 Before/After Outfit Tests

Virtual Try-On V2 takes a model photo plus garment images and generates a new on-model try-on shot. This post runs 6 before/after tests and shows the exact outputs.

The model tested: https://wiro.ai/models/wiro/virtual-try-on-v2

What Virtual Try-On V2 does

Virtual Try-On V2 targets fashion content pipelines. It accepts one human image and up to 13 garment images, then renders a new photo with the outfit applied. It can also generate video, but this test stays on images to keep comparisons clean.

• Inputs: inputImageHuman + inputImageClothes[]
• Controls used here: style, pose, plan, ratio, outputType
• Output: 1 image per run

Test setup

All 6 runs use the same human photo and the same 3 garment images. Only the shooting settings change.

Input: human

Input: garments

Before and after results (6 runs)

1) Keep original scene (virtual-try-on), standing, wide shot, 9:16

This run tries to preserve the original scene while swapping the outfit. Look for clean seams at the waist and realistic fabric folds near elbows.

2) Studio look, standing, wide shot, 9:16

Studio style pushes cleaner lighting and background control. This is usually the easiest setting for product teams because it matches catalog expectations.

3) Indoor lifestyle, standing, medium shot, 1:1

Medium shot trades full-body garment checks for a more product page friendly crop. It works best for tops, jackets, and accessories.

4) Outdoor location, standing, wide shot, 16:9

Wide landscape makes sense for homepage hero sections and lookbook banners. The risk is the outfit getting smaller in-frame.

5) Studio, side profile, wide shot, 9:16

Side profile can reveal fit issues fast. Hands, hair, and sleeve edges get harder here, so it is a good stress test.

6) Keep original scene (virtual-try-on), sitting, medium shot, 1:1

Sitting pose changes garment tension at the waist and knees. If the model handles this, it usually handles real-world catalog variety better.

Quick comparison

What it does well (and what to watch)

Works well for

• Generating fast outfit previews for PDP images and lookbooks
• Trying multiple backgrounds (studio vs indoor vs outdoor) without a new shoot
• Producing ratio variants for social and web (1:1, 9:16, 16:9)

Watch-outs

• Side profile and sitting poses can expose fabric warping and edge artifacts
• Busy garment photos can confuse layering when multiple items overlap
• Medium shots can hide lower-body items by design

Try it

Model link: https://wiro.ai/models/wiro/virtual-try-on-v2