KAMA: 3D Keypoint Aware Body Mesh Articulation

  • 2021-04-27 23:01:03
  • Umar Iqbal, Kevin Xie, Yunrong Guo, Jan Kautz, Pavlo Molchanov
  • 27

Abstract

We present KAMA, a 3D Keypoint Aware Mesh Articulation approach that allowsus to estimate a human body mesh from the positions of 3D body keypoints. Tothis end, we learn to estimate 3D positions of 26 body keypoints and propose ananalytical solution to articulate a parametric body model, SMPL, via a set ofstraightforward geometric transformations. Since keypoint estimation directlyrelies on image clues, our approach offers significantly better alignment toimage content when compared to state-of-the-art approaches. Our proposedapproach does not require any paired mesh annotations and is able to achievestate-of-the-art mesh fittings through 3D keypoint regression only. Results onthe challenging 3DPW and Human3.6M demonstrate that our approach yieldsstate-of-the-art body mesh fittings.

 

Quick Read (beta)

loading the full paper ...