AI Paper Atlas

Research map / 3D vision / Human pose estimation

Learning to Reconstruct 3D Human Pose and Shape via Model-Fitting in the Loop

Nikos Kolotouros, Georgios Pavlakos, Michael J. Black, Kostas Daniilidis

ICCV 2019 · 3D vision › Human pose estimation · approach: Human shape and pose regression · 1,090 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
add fitting in the training loop instead of static SMPLify fits (ablation)SPIN · 3DPWreconstruction error (mm): 66.3 -> 59.2Improved
add model-based supervision from static fits versus prior HMR-style 2D lossesSPIN static fits · 3DPWreconstruction error (mm): 70.2 -> 66.3Improved
add fitting in the training loop (ablation)SPIN · LSP foreground-backgroundaccuracy: 91.07 -> 91.83Improved
add fitting in the training loop (ablation)SPIN · LSP part segmentationaccuracy: 88.48 -> 89.41Improved
add fitting in the training loop (ablation)SPIN · LSP part segmentationf1: 0.65 -> 0.68Improved
add fitting in the training loop (ablation)SPIN · LSP foreground-backgroundf1: 0.86 -> 0.87Improved
use SPIN instead of HMR with 2D reprojection loss, with 3D ground truthSPIN · Human3.6Mreconstruction error (mm): 56.8 -> 41.1Improved
use SPIN instead of HMR, unpaired (no 3D ground truth)SPIN · Human3.6Mreconstruction error (mm): 66.5 -> 62.0Improved
use SPIN instead of HMR, unpairedSPIN · MPI-INF-3DHP absoluteMPJPE: 169.5 -> 124.8Improved
use SPIN instead of HMR, with 3D ground truthSPIN · MPI-INF-3DHP rigid alignmentMPJPE: 89.8 -> 67.5Improved
use SPIN instead of VNect, with 3D ground truthSPIN · MPI-INF-3DHP absolutePCK: 76.6 -> 76.4Hurt
use SPIN instead of SMPLify oracle at test timeSPIN · LSP foreground-backgroundaccuracy: 92.17 -> 91.83Hurt

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