search

sangyun884 / HR-VITON

Official PyTorch implementation for the paper High-Resolution Virtual Try-On with Misalignment and Occlusion-Handled Conditions (ECCV 2022).
831 stars 171 forks source link

Use my own cloth fail, but add your testdataset mask can #44

Closed dkhonker closed 1 year ago

dkhonker commented 1 year ago

Hello, I use my own clothes image to test but i get a bad result. img img

But I randomly chose a mask from your test set instead of my own cloth mask. I get a good result.

img

Do you know what the problem is？

maiskovich commented 1 year ago

@dkhonker what was the issue?

dkhonker commented 3 months ago

I found that the background of the data set was not (255,225,255), which affected the quality of the algorithm. The correct RGB of the background is (246,246,246).