Skip to content

Handle grayscale images on draw_bounding_boxes #4043

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Jun 11, 2021
Merged

Handle grayscale images on draw_bounding_boxes #4043

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
97226fe
allow single-channel-images in draw_bounding_boxes
sugatoray Jun 10, 2021
417acfd
Merge pull request #1 from sugatoray/sugatoray-patch-1
sugatoray Jun 10, 2021
d5cf449
Merge branch 'pytorch:master' into master
sugatoray Jun 10, 2021
8a384cb
change made for passing linting requirements
sugatoray Jun 10, 2021
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
4 changes: 4 additions & 0 deletions torchvision/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -180,6 +180,10 @@ def draw_bounding_boxes(
raise ValueError("Pass individual images, not batches")

ndarr = image.permute(1, 2, 0).numpy()
# allow single-channel-images
# shape: (1, H, W) with C = 1
Copy link
Member

Choose a reason for hiding this comment

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

Isn't the shape (H, W, C) instead, since we permuted the dimensions in the line above?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Yes, I realized the comment is not correct (only after the checks started running): it should be worded as below to avoid any confusion.

# image tensor shape: (1, H, W) with C = 1
# ndarr.shape: (H, W, C)

if ndarr.shape[-1] == 1:
ndarr = np.tile(ndarr, (1, 1, 3))
img_to_draw = Image.fromarray(ndarr)

img_boxes = boxes.to(torch.int64).tolist()
Expand Down