1 paper · 1 filter
Li Jin, Weikai Chen, Yujie Wang +7
Open-world promptable 3D semantic segmentation remains brittle as semantics are inferred in the input sensor coordinates. Yet, humans, in contrast, interpret parts via functional r…