AI Paper Atlas

Research map / 3D vision / Human pose estimation

Self-Supervised 3D Hand Pose Estimation From Monocular RGB via Contrastive Learning

Adrian Spurr, Aneesh Dahiya, Xi Wang, Xucong Zhang, Otmar Hilliges

ICCV 2021 · 3D vision › Human pose estimation · approach: Depth-based hand pose regression · 82 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 equivariant contrastive loss (PeCLR) instead of SimCLR NT-Xent (ablation)RN50 frozen encoder · FreiHAND3D EPE: 16.62 -> 16.05Improved
use equivariant contrastive loss (PeCLR) instead of SimCLR NT-Xent (ablation)RN50 frozen encoder · FreiHAND2D EPE: 12.05 -> 10.51Improved
use equivariant loss for translation augmentation (ablation)RN50 frozen encoder · FreiHAND3D EPE: 34% relative improvementImproved
use equivariant loss for rotation augmentation (ablation)RN50 frozen encoder · FreiHAND3D EPE: 57% relative improvementImproved
add PeCLR pre-training on FH and YT3D (ablation)RN50 · FreiHAND test3D PA-EPE: 0.83 -> 0.71Improved
add PeCLR pre-training on FH and YT3D (ablation)RN152 · FreiHAND test3D PA-EPE: 0.74 -> 0.66Improved
add PeCLR pre-training (ablation)RN152 · YouTube3DHands cross-dataset2D EPE: 22.1 -> 16.9Improved
add PeCLR pre-training (ablation)RN152 · YouTube3DHands cross-dataset3D PA-EPE: 3.05 -> 2.88Improved
add PeCLR pre-training (ablation)RN152 · FreiHAND cross-dataset eval3D EPE: 5.05 -> 4.56Improved
add unlabeled YT3D data to PeCLR pre-training (FH+YT3D vs FH) (ablation)RN152 · FreiHAND semi-supervised3D EPEImproved
add PeCLR pre-training with limited labels (ablation)RN152 · FreiHAND semi-supervised3D EPEImproved
add rotation augmentation with SimCLR (invariant loss) (ablation)SimCLR · FreiHAND3D EPEHurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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