AI Paper Atlas

Research map / 3D vision / Human pose estimation

A Pose-Only Geometric Constraint for Multi-Camera Pose Adjustment

Shunkun Liang, Banglei Guan, Bin Li, Qifeng Yu, Yang Shang

TPAMI 2026 · 3D vision › Human pose estimation · approach: Learned camera pose estimation · 5 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
replace bundle adjustment with pose-only constraint (MCPA vs MultiCol)multi-camera optimization · Synthetic (200 poses, 10000 points)runtime (s): 17.94 -> 7.11Improved
replace bundle adjustment with pose-only constraint (MCPA vs MultiCol)multi-camera optimization · Synthetic (200 poses, 10000 points)Hessian memory (MB): 95.13 -> 4.99Improved
replace bundle adjustment with pose-only constraint (MCPA vs MultiCol)multi-camera optimization · ETH3D Terrainsruntime (s): 56.85 -> 9.51Improved
add right-base residual as primary (MCPALR vs MCPA) (ablation)MCPA · Synthetic (50 poses, 1000 points)average rotation error (deg): 0.0181 -> 0.0128Improved
add right-base residual as primary (MCPALR vs MCPA) (ablation)MCPA · Synthetic (50 poses, 1000 points)runtime (s): 0.25 -> 0.36Hurt
use maximum-roundness base observation selection instead of random, maximum-disparity or maximum-theta (ablation)pose-only constraint · Synthetic Forward-Linear and Forward-Curvetriangulation errorImproved
triangulate from two well-selected base observations instead of all observations (ablation)3D point reconstruction · Synthetic Forward-Linear and Forward-Curvetriangulation errorImproved
use covariance-weighted statistically optimal triangulation (SOT) instead of NonLin or IRMP (ablation)triangulation · Synthetic, sigma_max = 10reconstruction error: improvement in precision by an order of magnitudeImproved
use pose-only methods under low image noiseMCPA/MCPALR vs MultiCol and BACS · Syntheticpose errorNo clear effect
use pose-only methods under high image noiseMCPA/MCPALR vs MultiCol and BACS · Syntheticpose errorImproved
use pose-only methods on ETH3DMCPALR vs MultiCol · ETH3D Electroaverage translation error (%): 4.029 -> 0.510Improved

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