AI Paper Atlas

Research map / 3D vision / Depth and stereo

Globally Optimal Vertical Direction Estimation in Atlanta World

Yinlong Liu, Guang Chen, Alois Knoll

TPAMI 2020 · 3D vision › Depth and stereo · approach: Camera model and calibration for SfM · 27 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
search vertical direction on S^2 instead of rotation search in SO(3)BnB vertical direction estimation · Synthetic Atlanta worldruntimeImproved
use exp-BnB or ste-circle-BnB bounds instead of ste-square or SCS bounds (ablation)BnB vertical direction estimation · Synthetic Atlanta worldruntimeImproved
use tighter ste-square or SCS bounds instead of exp or ste-circle bounds (ablation)BnB vertical direction estimation · Synthetic Atlanta worlditerationsImproved
use ste-square bounds instead of exp bounds (ablation)BnB vertical direction estimation · NYUv2median runtime (s): 0.134 -> 4.256Hurt
use ste-square bounds instead of exp bounds (ablation)BnB vertical direction estimation · NYUv2median iterations: 816 -> 478Improved
use ste-square bounds instead of ste-circle bounds (ablation)BnB vertical direction estimation · Outdoor Bremenmedian runtime (s): 1.182 -> 88.470Hurt
use SCS parametrization instead of stereographic square parametrization (ablation)BnB vertical direction estimation · Outdoor Bremenmedian runtime (s): 88.470 -> 153.356Hurt
increase RANSAC outlier proportion assumption from 0.65 to 0.95 (ablation)RANSAC · NYUv2median runtime (s): 0.007 -> 0.344Hurt
use sequential vertical-then-horizontal estimation instead of rotation searchManhattan frame estimation · Synthetic Manhattan worldaccuracy at large noiseHurt
use sequential vertical-then-horizontal estimation instead of rotation searchManhattan frame estimation · Synthetic Manhattan worldruntimeImproved
use exp-BnB instead of other S^2 bounds under high outlier ratio (ablation)BnB vertical direction estimation · Synthetic high outlier / large noiseruntimeImproved

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