AI Paper Atlas

Research map / 3D vision / Depth and stereo

Secrets of Event-Based Optical Flow

Shintaro Shiba, Yoshimitsu Aoki, Guillermo Gallego

ECCV 2022 · 3D vision › Depth and stereo · approach: Learned optical and scene flow · 120 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 multi-reference focus loss instead of single-reference loss (ablation)tile-based contrast maximization · MVSEC/DSEC (qualitative)IWE sharpness / flow regularityImproved
add time-aware flow (Burgers') over no time awareness (ablation)Ours · ECD slider_depthFWL: 1.882863 -> 1.9298641427098424Improved
add time-aware flow (Burgers') over no time awareness (ablation)Ours · DSEC thun_00_aFWL: 1.39 -> 1.42Improved
add time-aware flow on MVSEC (ablation)Ours · MVSEC indoor_flying1 (dt=1)AEE: 0.42152 -> 0.42281No clear effect
add time-aware flow on MVSEC FWL (ablation)Ours · MVSEC dt=4FWL: 1.17 -> 1.17No clear effect
use multi-scale (five scales) instead of single finest scale (ablation)Ours · MVSEC (qualitative)flow regularityImproved
use three scales instead of five (ablation)Ours · MVSEC (qualitative)flow regularityHurt
use image variance loss instead of gradient magnitude (ablation)Ours · MVSEC indoor_flying1 (dt=4)AEE: 1.68217 -> 1.69705397Hurt
use average timestamp loss instead of gradient magnitude (ablation)Ours · MVSEC (dt=4)AEE: 1.68217 -> >99Hurt
increase regularizer weight from 0.0025 to 0.025 (ablation)Ours · MVSEC outdoor_day1 (dt=4)AEE: 1.24740 -> 1.85784597Hurt
increase regularizer weight from 0.0025 to 0.025 (ablation)Ours · MVSEC indoor_flying1 (dt=4)AEE: 1.68217 -> 1.517299916933Improved
increase regularizer weight from 0.0025 to 0.25 (ablation)Ours · MVSEC indoor_flying1 (dt=4)AEE: 1.68217 -> 1.885701460289Hurt

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