Sapiens Body

Meta open foundation models for human vision — pose, depth, segmentation, normal.