1 paper
Xiangyu Xu, Lijuan Liu, Shuicheng Yan
Existing Transformers for monocular 3D human shape and pose estimation typically have a quadratic computation and memory complexity with respect to the feature length, which hinder…