方法

MFRM means ‘multi-modal feature refinement mechanism’，MMAFM means ‘multi-modal and multi-scale attention fusion model’，RPM means ‘residual prediction module’

scale attention weights U${}_R^S$,U${}_D^S$



enhanced RGB and depth features (i.e.,F${}_{ER}^i$,F${}_{ED}^i$)，HGAM means ‘high-level guided attention mechanism’

作者未提供代码