AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

FlowFormer: A Transformer Architecture for Optical Flow

Zhaoyang Huang, Xiaoyu Shi, Chao Zhang, Qiang Wang, Ka Chun Cheung, Hongwei Qin, Jifeng Dai, Hongsheng Li

ECCV 2022 · Architectures and efficiency › RNNs and normalization · approach: Dataset condensation and flow estimation · 430 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
replace RAFT multi-level cost retrieval decoder with latent cost tokens plus cost memory decoder (K=8, D=128) (ablation)FlowFormer (RAFT-based ablation) · Sintel train cleanAEPE: 1.53 -> 1.44Improved
increase latent tokens from K=4,D=32 to K=8,D=128 (ablation)FlowFormer (RAFT-based ablation) · Sintel train cleanAEPE: 1.66 -> 1.44Improved
replace CNN image encoder with Twins trained from scratch (ablation)FlowFormer (RAFT-based ablation) · Sintel train finalAEPE: 2.80 -> 2.86Hurt
use ImageNet-pretrained Twins instead of CNN image encoder (ablation)FlowFormer (RAFT-based ablation) · Sintel train cleanAEPE: 1.44 -> 1.29Improved
use ImageNet-pretrained Twins instead of CNN image encoder (ablation)FlowFormer (RAFT-based ablation) · KITTI-15 trainF1-all (%): 17.64 -> 16.16Improved
add intra-cost-map attention only (ablation)FlowFormer (RAFT-based ablation) · Sintel train finalAEPE: 2.72 -> 2.89Hurt
add one AGT layer (intra + inter attention) (ablation)FlowFormer (RAFT-based ablation) · Sintel train cleanAEPE: 1.29 -> 1.20Improved
stack three AGT layers instead of none (ablation)FlowFormer (RAFT-based ablation) · Sintel train finalAEPE: 2.72 -> 2.57Improved
stack three AGT layers instead of none (ablation)FlowFormer (RAFT-based ablation) · KITTI-15 trainF1-all (%): 16.16 -> 15.15Improved
use FlowFormer (small) with GMA CNN encoder versus GMA (ablation)FlowFormer (RAFT-based ablation) · Sintel train cleanAEPE: 1.30 -> 1.20Improved
enlarge GMA to GMA-L (17.0M parameters) to match sizeFlowFormer (RAFT-based ablation) · Sintel train cleanAEPE: 1.30 -> 1.33No clear effect
give GMA a pretrained Twins encoderFlowFormer (RAFT-based ablation) · KITTI-15 trainF1-epe: 4.69 -> 4.98Hurt

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