Skip to content
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.

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

For obb detection, the corner of the box is not on the image #502

Closed
DWendou opened this issue Jul 5, 2024 · 5 comments
Closed

For obb detection, the corner of the box is not on the image #502

DWendou opened this issue Jul 5, 2024 · 5 comments
Labels
Implemented question Further information is requested

Comments

@DWendou
Copy link

DWendou commented Jul 5, 2024

For the box at the edge of the image, some box corners may not be on the image, how to deal with this? thank you

@DWendou DWendou changed the title the corner of the box is not on the image For obb detection, the corner of the box is not on the image Jul 5, 2024
@CVHub520
Copy link
Owner

CVHub520 commented Jul 5, 2024

For the box at the edge of the image, some box corners may not be on the image, how to deal with this? thank you

When exporting the image with bounding boxes, any box that is partially or completely outside the image boundaries will be automatically ignored. This ensures that only relevant annotations within the visible image area are considered. If the full object is not within the image frame, it will not be included in the exported data.

Additionally, if you require a more refined approach, you have the option to apply a mask to the image. This mask can be used to specifically define the areas of interest, allowing you to include or exclude certain parts of the bounding boxes that extend beyond the image edges. Currently, you should implement it by manully.

Currently, this feature is not automated, and you would need to implement it manually. This means you would have to create and apply the mask yourself according to your specific requirements.

@CVHub520 CVHub520 added the question Further information is requested label Jul 5, 2024
@DWendou
Copy link
Author

DWendou commented Jul 5, 2024

ww
Are you saying you don't want any goals like that? Or I can make a mask manually

@CVHub520
Copy link
Owner

CVHub520 commented Jul 5, 2024

Sorry for the wait. Specifically, you can consider the following strategies:

  1. For targets that slightly exceed the image boundaries, you can try to shrink the bounding box. Although this may not completely cover the target, it can ensure the legality of the box.
  2. For targets that significantly exceed the image boundaries, you can choose to ignore them.
  3. For other situations, you can use masking techniques to mitigate these effects.

@CVHub520
Copy link
Owner

Hi, @DWendou:

Thank you for raising the issue regarding handling boxes at the edge of the image. I’ve addressed this and now support annotating rotated boxes that may extend beyond the image boundaries.

Please update to the latest source code and give it a try. If you encounter any further issues or need additional adjustments, feel free to let us know.

#545

@DWendou
Copy link
Author

DWendou commented Jul 25, 2024

Hi, @DWendou:

Thank you for raising the issue regarding handling boxes at the edge of the image. I’ve addressed this and now support annotating rotated boxes that may extend beyond the image boundaries.

Please update to the latest source code and give it a try. If you encounter any further issues or need additional adjustments, feel free to let us know.

#545

That's great. Thank you so much for your work. I'll try again later

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Implemented question Further information is requested
Projects
None yet
Development

No branches or pull requests

2 participants