Paper-Conference

MICDrop: Masking Image and Depth Features via Complementary Dropout for Domain‑Adaptive Semantic Segmentation