AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Self-Supervised Transfer Learning for Hand Mesh Recovery From Binocular Images

Zheng Chen, Sihan Wang, Yi Sun, Xiaohong Ma

ICCV 2021 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Pretext-task visual pre-training · 11 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 self-supervised binocular training on STB to the FreiHAND-pretrained modelhand mesh model · STBmean joint error (mm): 44.52mm to 11.14mmImproved
add photo-consistency loss (ablation)self-supervised hand mesh framework · STBmean joint error (mm): 25.36 -> 11.14Improved
add edge loss (ablation)self-supervised hand mesh framework · STBmean joint error (mm): 13.46 -> 11.14Improved
add Laplacian loss (ablation)self-supervised hand mesh framework · STBmean joint error (mm): 11.36 -> 11.14Improved
add mask loss (ablation)self-supervised hand mesh framework · STBmean joint error (mm): 11.23 -> 11.14Improved
add Chamfer distance loss (ablation)self-supervised hand mesh framework · STBmean joint error (mm): 11.37 -> 11.14Improved
add normal loss (ablation)self-supervised hand mesh framework · STBmean joint error (mm): 11.23 -> 11.14Improved
use binocular disparity for z instead of regressing z per branch (monocular)self-supervised hand mesh framework · STBmean joint error (mm): 12.66 -> 11.14Improved

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