AI Paper Atlas

Research map / 3D vision / Face and shape

DEMEA: Deep Mesh Autoencoders for Non-Rigidly Deforming Objects

Edgar Tretschk, Ayush Tewari, Michael Zollhöfer, Vladislav Golyanik, Christian Theobalt

ECCV 2020 · 3D vision › Face and shape · approach: Articulated and non-rigid shape learning · 43 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use regressed EDL rotations instead of graph loss (GL) (ablation)DEMEA · DFaust (cm), latent dim 8average per-vertex error: 8.92 -> 6.69Improved
use regressed EDL rotations instead of graph loss (GL) (ablation)DEMEA · CoMA (mm), latent dim 32average per-vertex error: 0.99 -> 0.81Improved
use regressed EDL rotations instead of local Procrustes (LP) (ablation)DEMEA · DFaust (cm), latent dim 8average per-vertex error: 7.71 -> 6.69Improved
use regressed EDL rotations instead of local Procrustes (LP) (ablation)DEMEA · Textureless cloth (mm), latent dim 32average per-vertex error: 7.96 -> 6.40Improved
use regressed EDL rotations instead of local Procrustes (LP) (ablation)DEMEA · DFaust (cm), latent dim 32average per-vertex error: 2.22 -> 2.23No clear effect
add embedded deformation layer instead of learned upsampling (CA) (ablation)DEMEA · SynHand5M (mm), latent dim 32average per-vertex error: 2.60 -> 2.51Improved
add embedded deformation layer instead of learned upsampling (CA) (ablation)DEMEA · DFaust (cm), latent dim 32average per-vertex error: 2.07 -> 2.23Hurt
add embedded deformation layer instead of learned upsampling (CA) (ablation)DEMEA · CoMA (mm), latent dim 32average per-vertex error: 0.72 -> 0.81Hurt
use graph-convolutional decoder with EDL instead of almost-linear FC network (FCA) (ablation)DEMEA · SynHand5M (mm), latent dim 8average per-vertex error: 15.10 -> 8.12Improved
use graph-convolutional decoder with EDL instead of almost-linear FC network (FCA) (ablation)DEMEA · Textureless cloth (mm), latent dim 8average per-vertex error: 15.63 -> 11.28Improved
use spiral convolutions instead of spectral convolutions (ablation)DEMEA · SynHand5M (mm), latent dim 32average per-vertex error: 3.83 -> 2.51Improved
use spiral convolutions instead of spectral convolutions (ablation)DEMEA · DFaust, latent dim 8average per-vertex error: 6.56 -> 6.69Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy