How do handle overlapping segmentation masks?

In maskrcnn I’m pretty sure it’s valid to output overlapping segmentation masks, but I’m not sure how to handle this in wandb Masks and I haven’t been able to find an example. The documentation says each “Mask” is a 2D numpy array filled with class ids, which implies one class per pixel, right?

Is there a way to visualize the masks if they’re overlapping?

I was told by wandb support this was not possible, posting here for transparency

1 Like

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.