deocclusion icon indicating copy to clipboard operation
deocclusion copied to clipboard

Published 20 hours ago • verified publisher icon XiaohangZhan

Inference result is not expected, totally fail

Open weihaosky opened this issue 3 years ago • 3 comments

Hi, I want to use this code to complete the human body. The input mask is extracted with mask-rcnn. But the result is not expected: image I also tried the example image in the repo, but the result is also bad: image What is the problem? Many thanks!

weihaosky avatar May 11 '21 12:05 weihaosky

  1. The mask is too noisy.
  2. There should no margin between two occluding masks, adjust parameter dilate_kernel to avoid it.

XiaohangZhan avatar May 14 '21 08:05 XiaohangZhan

Hi, I want to use this code to complete the human body. The input mask is extracted with mask-rcnn. But the result is not expected: image I also tried the example image in the repo, but the result is also bad: image What is the problem? Many thanks!

I now want to experiment with my own data, but I don't know how to label it. Can you help me ?

wmrenr avatar May 30 '22 11:05 wmrenr

Hi, I want to use this code to complete the human body. The input mask is extracted with mask-rcnn. But the result is not expected: image I also tried the example image in the repo, but the result is also bad: image What is the problem? Many thanks!

I now want to experiment with my own data, but I don't know how to label it. Can you help me ?

I now want to experiment with my own data, but I don't know how to label it. Can you help me ?

quanw8781 avatar Sep 22 '22 01:09 quanw8781