Looks like Mask3D ([https://jonasschult.github.io/Mask3D/] is the current best performing model for 3D semantic instance segmentation. Is there any plan for implementing it in mmdetection3d?
Hi, thanks for your proposal. We'll evaluate its workload and consider it seriously. We'll appreciate it if you can create a PR to support this method.
Looks like Mask3D ([https://jonasschult.github.io/Mask3D/] is the current best performing model for 3D semantic instance segmentation. Is there any plan for implementing it in mmdetection3d?