AI Paper Atlas

Research map / 3D vision / Depth and stereo

Uncertainty Estimates and Multi-Hypotheses Networks for Optical Flow

Eddy Ilg, Ozgun Cicek, Silvio Galesso, Aaron Klein, Osama Makansi, Frank Hutter, Thomas Brox

ECCV 2018 · 3D vision › Depth and stereo · approach: Learned optical and scene flow · 232 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
train with predictive negative log-likelihood loss instead of EPE loss (Dropout) (ablation)Dropout ensemble · Sintel train cleanAUSE: 0.212 -> 0.158Improved
train with predictive loss instead of EPE loss (Bootstrapped Ensemble) (ablation)BootstrappedEnsemble · Sintel train cleanAUSE: 0.209 -> 0.127Improved
train with predictive loss instead of EPE loss (SGDR Ensemble) (ablation)SGDREnsemble · Sintel train cleanAUSE: 0.191 -> 0.134Improved
use predictive instead of empirical training (ablation)Ensembles · Sintel train cleanEPE: 3.41 -> 3.46Hurt
add a merging network on top of Bootstrapped Ensemble (ablation)BootstrappedEnsemble (Pred) · Sintel train cleanAUSE: 0.127 -> 0.102Improved
add a merging network on top of Bootstrapped Ensemble (ablation)BootstrappedEnsemble (Pred) · Sintel train cleanEPE: 3.46 -> 3.20Improved
replace Bootstrapped Ensemble with multi-hypothesis FlowNetH plus merging networkBootstrappedEnsemble-Merged · Sintel train cleanruntime: 332ms -> 60msImproved
replace Bootstrapped Ensemble with FlowNetH plus merging networkBootstrappedEnsemble-Merged · Sintel train cleanAUSE: 0.102 -> 0.095No clear effect
use ensemble of 8 instead of single predictive network (ablation)FlowNetC (Pred) · Sintel train cleanAUSE: 0.133 -> 0.134No clear effect
use Dropout instead of other ensemble typespredictive ensembles · Sintel train cleanAUSE: 0.158 vs 0.127Hurt
stack one refinement network (S) on FlowNetH-Pred-Merged (ablation)FlowNetH-Pred-Merged · Sintel train cleanAUSE: 0.117 -> 0.091Improved
stack two refinement networks (SS) on FlowNetH-Pred-Merged (ablation)FlowNetH-Pred-Merged · KITTI 2012+2015 validation splitEPE: 7.84 -> 6.50Improved

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