search

switchablenorms / CelebAMask-HQ

A large-scale face dataset for face parsing, recognition, generation and editing.
2.13k stars 352 forks source link

g_mask.py generated masks are all blank? #88

Open charliewang789 opened 3 years ago

charliewang789 commented 3 years ago

could you please let me know why the mask files generated are all blank?

changrannn commented 3 years ago

Hi! I also meet this question, I want to ask if you have solved this problem？ Thank you very much!

greecehalf commented 3 years ago

I met the same issue. Maybe some nuances in the path that you have to modify.

tien1504 commented 3 years ago

Me too, any update on this. I noticed that the CelebAMask-HQ-mask-anno folder just contains few folders (0, 1, 10, 11, 12, 13) while the g_mask requires 0->14 folders. Where are the others?

tien1504 commented 3 years ago

I figured it out. We need to change the folder name 'CelebAMaskHQ-mask-anno' to 'CelebAMask-HQ-mask-anno' in the g_mask.py to match with the folder name in the dataset first.

SahadevPoudel commented 1 year ago

Anyone facing this problem, can do this simple step:

cv2.imwrite(filename_save, im_base) TO cv2.imwrite(filename_save, im_base*255)