AI Paper Atlas

Research map / Video / Action recognition

ActionFormer: Localizing Moments of Actions with Transformers

Chen-Lin Zhang, Jianxin Wu, Yin Li

ECCV 2022 · Video › Action recognition · approach: Disentangled action recognition · 525 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 convolutional backbone with Transformer (ablation)AF Base · THUMOS14average mAP: 46.6 -> 62.7Improved
add layer norm in heads (ablation)ActionFormer · THUMOS14average mAP: 62.7 -> 65.4Improved
add center sampling (ablation)ActionFormer · THUMOS14average mAP: 65.4 -> 66.8Improved
add position encoding (ablation)ActionFormer · THUMOS14average mAP: 66.8 -> 66.7No clear effect
add score fusion (ablation)ActionFormer · THUMOS14average mAP: 66.8 -> 62.1Hurt
add multiscale feature pyramid (vs single level) (ablation)ActionFormer · THUMOS14average mAP: 47.6 -> 66.8Improved
use local window 19 instead of global attention (ablation)ActionFormer · THUMOS14average mAP: 66.8 -> 66.8No clear effect
initial regression range [0,4) vs [0,16) (ablation)ActionFormer (3 levels) · THUMOS14average mAP: 50.2 -> 60.1Improved
increase pyramid levels from 3 to 6 (ablation)ActionFormer · THUMOS14average mAP: 60.1 -> 66.8Improved
increase pyramid levels from 6 to 7 (ablation)ActionFormer · THUMOS14average mAP: 66.8 -> 66.2No clear effect
vary regression loss weight lambda_reg from 1 (ablation)ActionFormer · THUMOS14average mAP: 65.4 to 66.8No clear effect
increase max training length from 576 to 2304 (ablation)ActionFormer · THUMOS14average mAP: 65.7 -> 66.8Improved

Later papers that compared with this method

21 later accepted papers report ActionFormer as a baseline in their results tables; 21 of them report a better number on most of the results they share with it.

13 more not listed.

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