AI Paper Atlas

Research map / 3D vision / Human pose estimation

Point-to-Point Regression PointNet for 3D Hand Pose Estimation

Liuhao Ge, Zhou Ren, Junsong Yuan

ECCV 2018 · 3D vision › Human pose estimation · approach: Depth-based hand pose regression · 179 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
stack two PointNet modules instead of one (ablation)P2P Reg. PointNet · NYUmean error distance: 10.3mm -> 9.5mmImproved
output unit vector fields instead of offset fields (ablation)two-stacked P2P Reg. · NYUmean error distance: 10.8mm -> 9.5mmImproved
use point-to-point regression instead of direct regressiontwo-stacked PointNet · NYUmean error distance: 10.8mm -> 9.5mmImproved
add direct-regression replacement for divergent joints (ablation)two-stacked P2P Reg. · NYUmean error distance: 9.5mm -> 9.2mmImproved
add PCA constraint on the pose (ablation)P2P Reg.+Direct Reg. · NYUmean error distance: 9.2mm -> 9.1mmImproved
increase candidate points M from 25 to 55 (weighted average) (ablation)P2P Reg. inference · NYUmean error distance: 9.46mm -> 9.71mmHurt
use weighted instead of unweighted average at M=55 (ablation)P2P Reg. inference · NYUmean error distance: 9.84mm -> 9.71mmImproved
use weighted instead of unweighted average at M=5 (ablation)P2P Reg. inference · NYUmean error distance: 9.50mm -> 9.50mmNo clear effect
use P2P Reg. instead of DenseReghand pose estimation · NYUmean error distance: 10.2mm -> 9.1mmImproved
use P2P Reg. instead of DenseReghand pose estimation · ICVLmean error distance: 7.3mm -> 6.3mmImproved
use P2P Reg. instead of DenseReghand pose estimation · MSRAmean error distance: 7.2mm -> 7.7mmHurt

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