Skip to content
This repository has been archived by the owner on Jul 2, 2021. It is now read-only.

Return rgb when rgba is loaded #734

Merged
merged 15 commits into from
Feb 6, 2019
Merged

Return rgb when rgba is loaded #734

Changes from 1 commit
Commits
Show all changes
15 commits
Select commit Hold shift + click to select a range
649a886
return rgb when rgba is loaded
yuyu2172 Nov 18, 2018
f707028
fix typo
yuyu2172 Nov 26, 2018
213a3e7
add alpha option
yuyu2172 Feb 2, 2019
08d736b
use cv2.IMREAD_UNCHANGED when alpha is not None
yuyu2172 Feb 3, 2019
0ab431d
fix bug
yuyu2172 Feb 3, 2019
b7b5cbc
test
yuyu2172 Feb 3, 2019
ad09b22
make test more strict
yuyu2172 Feb 3, 2019
c2335c9
flake8
yuyu2172 Feb 3, 2019
b8bbb51
fix bug
yuyu2172 Feb 3, 2019
2aaa9ed
merge master
yuyu2172 Feb 3, 2019
0f38d43
fix doc
yuyu2172 Feb 3, 2019
8cd06d0
support python2
yuyu2172 Feb 3, 2019
d586772
fix comment
yuyu2172 Feb 3, 2019
11aecce
fix typo
yuyu2172 Feb 6, 2019
b8b71d8
merge
yuyu2172 Feb 6, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions chainercv/utils/image/read_image.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,11 @@ def _read_image_cv2(path, dtype, color):
# reshape (H, W) -> (1, H, W)
return img[np.newaxis].astype(dtype)
else:
# alpha channel is inclued
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

included?

if img.shape[-1] == 4:
bgr = img[:, :, :3]
alpha = img[:, :, 3:] / 255
img = bgr * alpha + 255 * np.ones_like(bgr) * (1 - alpha)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think there are three options.

  1. just ignore alpha channel
  2. blend with black image
  3. blend with white image

Why do you choose 3.? Personally, I prefer 1.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is because the alpha channel contains meaningful information (for some part, the value is in (0, 1)).

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  1. is not acceptable?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is 2 better?
I thought 3 was more natural

img = img[:, :, ::-1] # BGR -> RGB
img = img.transpose((2, 0, 1)) # HWC -> CHW
return img.astype(dtype)
Expand Down