multi - channel images support #10381
Description
Actions before raising this issue
- I searched the existing issues and did not find anything similar.
- I read/searched the docs
Is your feature request related to a problem? Please describe.
I'm training segmentation models and use cvat for creating polygon labels in my images. However, sometimes it is hard to see the region that needs segmenting. To solve these issues, my images always have a twin that is created from an alternative sensor and provides additional information. If I would stack them I would get a 4-channel image (e.g. RGB + IR)
So far I couldn't find any way to upload these paired images and than switch quickly between them. cvat only has the context images which isn't the same. Annotation is only done in the primary image.
Describe the solution you'd like
One could use the context-images (related_images folder) to upload the twin but instead of showing them in an image stripe below or next to the image, one could switch the display of the current image by pressing a button. Zoom state and region in the view is maintained but the other channel is shown. Labels should still be visible so that the label could be edited with the help of the context image.
labelbox or DarwinV7 provide this functionality, for example.
Describe alternatives you've considered
No response
Additional context
No response