AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

MegaFlow: Zero-Shot Large Displacement Optical Flow

Dingxi Zhang, Fangjinhua Wang, Marc Pollefeys, Haofei Xu

ECCV 2026 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Transductive and semantic-dictionary zero-shot · 6 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 VGGT pretrained prior to a 12-layer backbone (ablation)MegaFlow · Zero-shot Sintel FinalEPE: 3.31 -> 2.23Improved
scale transformer from 6 to 12 layers without pre-training (ablation)MegaFlow from scratch · Zero-shot Sintel CleanEPE: 2.05 -> 2.13Hurt
keep transformer trainable instead of frozen (ablation)MegaFlow · Zero-shot KITTIFl-all: 14.7 -> 10.9Improved
add CNN feature fusion (ablation)MegaFlow · Zero-shot KITTIFl-epe: 4.71 -> 3.00Improved
add temporal attention with 4-frame input (ablation)MegaFlow · Zero-shot Sintel FinalEPE: 1.99 -> 1.83Improved
increase input frames from 2 to 4 with temporal attention (ablation)MegaFlow · Zero-shot Sintel FinalEPE: 2.08 -> 1.83Improved
increase input frames from 2 to 6 with temporal attention (ablation)MegaFlow · Zero-shot KITTIFl-epe: 3.00 -> 3.67Hurt
add TartanAir training stage (ablation)MegaFlow · Zero-shot Sintel CleanEPE: 0.94 -> 0.85Improved
add 2-frame FlyingThings warmup stage (ablation)MegaFlow · Zero-shot KITTIFl-all: 12.2 -> 11.1Improved
increase refinement iterations from 1 to 8 (ablation)MegaFlow · Zero-shot KITTIFl-all: 19.03 -> 11.14Improved
increase refinement iterations beyond 8 (K=10) (ablation)MegaFlow · Zero-shot Sintel FinalEPE: 1.83 -> 1.85No clear effect
fine-tune on Kubric after flow-mix trainingMegaFlow · TAP-Vid meandelta_avg: 73.6 -> 79.6Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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