AI Paper Atlas

Research map / 3D vision / Depth and stereo

Event-Intensity Stereo: Estimating Depth by the Best of Both Worlds

Mohammad Mostafavi, Kuk-Jin Yoon, Jonghyun Choi

ICCV 2021 · 3D vision › Depth and stereo · approach: Learned stereo with specialized sensors · 46 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
add intensity frames to event-only stereostereo network · MVSEC split 2mean depth error (cm): 25.18 -> 18.43Improved
add intensity frames to event-only stereostereo network · ECARmean depth error (cm): 22.3 -> 11.8Improved
add intensity frames to event-only stereostereo network · MVSEC split 1mean depth error (cm): 13.27 -> 13.74Hurt
add events to intensity-only stereostereo network · MVSEC split 3mean depth error (cm): 23.78 -> 22.36Improved
use EIS instead of the events-only baselinestereo depth · DSECMAE: 0.576 -> 0.396Improved
use event-only ES variant instead of the events-only baselinestereo depth · DSECMAE: 0.576 -> 0.529Improved
add feature pyramid network (ablation)full network · ECAR subsetmean depth error (cm): 37.8 -> 8.3Improved
add deformable aggregation network (ablation)full network · ECAR subsetmean depth error (cm): 23.4 -> 8.3Improved
add multi-scale refinement (ablation)full network · ECAR subsetmean depth error (cm): 13.8 -> 8.3Improved
train with missing modalitiesEIS with no left event stack · missing data testmean depth error (cm): 87.5 -> 17.9Improved
train with missing modalitiesEIS with full modality at test · missing data testmean depth error (cm): 10.6 -> 16.8Hurt
use EIS instead of IS under inconsistent left-right lightingstereo depth · inconsistent lightmean depth error (cm): 69.0 -> 32.0Improved

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