Nowadays I dont see a lot of papers addressing 2D semantic segmentation problem statements be it supervised, semi-supervised, domain adaptation. Is the problem statement saturated? Are there any promising research directions in segmentation except open-set segmentation?
[link] [comments]