AI Paper Atlas

Research map / 3D vision / Human pose estimation

Camera-Space Hand Mesh Recovery via Semantic Aggregation and Adaptive 2D-1D Registration

Xingyu Chen, Yufeng Liu, Chongyang Ma, Jianlong Chang, Huayan Wang, Tian Chen, Xiaoyan Guo, Pengfei Wan, Wen Zheng

CVPR 2021 · 3D vision › Human pose estimation · approach: Human mesh recovery from images · 87 citations (OpenAlex)

Check 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
replace the baseline decoder with the Inception Spiral Module-based spiral decoder (ablation)YoutubeHand · FreiHANDPA-MPVPE: 9.06 -> 8.54Improved
use grouped semantic landmark heatmaps instead of separate landmark heatmaps (ablation)root-relative mesh recovery · FreiHANDPA-MPVPE: 7.77 -> 7.60Improved
concatenate silhouette and landmark heatmaps (ablation)root-relative mesh recovery · FreiHANDPA-MPVPE: 7.77 -> 7.95Hurt
sum landmark heatmaps into a single heatmap, discarding individual joint semantics (ablation)root-relative mesh recovery · FreiHANDPA-MPVPE: 7.77 -> 7.94Hurt
add a second stack with grouped landmark cues (ablation)group(Hp) mesh recovery · FreiHANDPA-MPVPE: 7.60 -> 7.46Improved
add part-based semantic grouping (ablation)joint landmark aggregation · FreiHANDPA-MPVPE: 7.77 -> 7.65Improved
add level-based semantic grouping (ablation)joint landmark aggregation · FreiHANDPA-MPVPE: 7.77 -> 7.69Improved
add tip-based semantic grouping (ablation)joint landmark aggregation · FreiHANDPA-MPVPE: 7.77 -> 7.65Improved
combine 2D landmark registration with 1D silhouette-span registration and adaptive fusion (ablation)CMR-P · RHD root recoveryCS-MPVPE: 53.7 -> 52.7Improved
use 12 projection axes rather than 2 in adaptive 2D-1D registration (ablation)CMR-P · RHD root recoveryCS-MPVPE: 52.7 -> 52.1Improved
use PnP instead of the proposed registrationCMR-P root recovery · RHD root recoveryCS-MPVPE: 99.0 vs. 52.1Hurt

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