Modality adaptation via feature difference learning for depth human parsing

Shaofei Huang, Tianrui Hui, Yue Gong, Fengguang Peng, Yuqiang Fang, Jingwei Wang, Bin Ma, Xiaoming Wei, Jizhong Han

Abstract: Highlights•An MAFDL pipeline leveraging RGB semantic knowledge to enhance depth human parsing.•DGDA to bridge the RGB-depth modality gap by learning inter-modal feature difference.•FAC as explicit supervision at pixel and batch levels for depth feature adaptation.•State-of-the-art performance on the NTURGBD-Parsing-4K dataset.