1 paper
Wei Wang, Yiding Sun, Yuyan Wang +4
Point cloud video representation learning is crucial for 3D dynamic scene understanding. In this paper, we propose MoSaiC, a novel Motion-Saliency Complementary masked modeling fra…