Two semantic segmentations from a single Perception Camera

Is there are a way to get two separate segmentation passes from a single frame?
Ideally I would just add a new Camera Labeler with two different label config files, but currently this throws an error.

 

Can you share the code that you are using that throws the error? We will look into this.

 

No code involved. When I add two different label config files to the perception cam I get this error:
Semantic segmentation was run twice in the same frame. Multiple semantic segmentations are not currently supported.

 

Andrezej, multiple instances of semantic segmentation is high on our list, to be done in the next release. We'll update this thread when we have a PR open, if you would like to take an early look.

 

Thanks Jon,
Do you have a timeline for this?
Or should I try to figure out a workaround for this?
using two perception cameras, replacing annotation files during runtime and changing the scene every other frame and a bunch of other things don't really work, so I think that would require some kind of a custom pass, similar to instance segmentation, that would work with HDRP. Not sure if there's an example implementation like that.

 

Haven't tried it yet, but I guess since the new update has instance segmentation I can use this as a workaround.

 

Locking this thread for inactivity. If you need help, please email simulation-help@unity3d.com.