AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Conditional Prior Networks for Optical Flow

Yanchao Yang, Stefano Soatto

ECCV 2018 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Pretext-task visual pre-training · 34 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 FlowNetS with learned conditional prior, unsupervised, on Flying Chairs (CPNFlow-C vs FlowNetS)FlowNetS · KITTI 2012 TrainEPE: 8.26 -> 7.33Improved
train with conditional prior on Flying Chairs instead of supervised FlowNetCFlowNetC · KITTI 2012 TrainEPE: 9.35 -> 7.33Improved
train on KITTI with the conditional prior (CPNFlow-K) vs supervised FlowNet2FlowNet2 · KITTI 2012 TrainEPE: 4.09 -> 3.76Improved
train on real KITTI videos (CPNFlow-K) instead of Flying Chairs (CPNFlow-C)CPNFlow · KITTI 2012 TrainEPE: 7.33 -> 3.76Improved
train on KITTI instead of Flying ChairsCPNFlow · Sintel Train cleanEPE: 4.87 -> 6.46Hurt
add occlusion reasoning to CPNFlow-K (ablation)CPNFlow-K · KITTI 2012 TrainEPE: 3.76 -> 3.11Improved
add occlusion reasoning to CPNFlow-K (ablation)CPNFlow-K · KITTI 2015 TrainEPE: 9.63 -> 7.82Improved
add occlusion reasoning to CPNFlow-K (ablation)CPNFlow-K · Sintel Train finalEPE: 7.12 -> 7.52Hurt
use learned conditional prior (CPNFlow-K) instead of UnFlowC unsupervised lossUnFlowC · KITTI 2012 TrainEPE: 7.11 -> 3.76Improved
use learned prior with occlusion (CPNFlow-K-o) vs UnFlowC-ocUnFlowC-oc · KITTI 2012 TrainEPE: 3.78 -> 3.11Improved
use learned prior (CPNFlow-C) instead of LDOF variational regularizerLDOF · KITTI 2012 TrainEPE: 13.7 -> 7.33Improved
use learned prior CPNFlow-C vs supervised FlowNetS on Sintel cleanFlowNetS · Sintel Train cleanEPE: 4.50 -> 4.87No clear effect

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