AI Paper Atlas

Research map / Video / Action recognition

Spatiotemporal Self-Attention Modeling with Temporal Patch Shift for Action Recognition

Wangmeng Xiang, Chao Li, Biao Wang, Xihan Wei, Xian-Sheng Hua, Lei Zhang

ECCV 2022 · Video › Action recognition · approach: Spatiotemporal CNNs for action recognition · 51 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 temporal patch shift with a single center patch (ablation)Swin-T · Something-Something V1top-1: 40.6 -> 45.3Improved
shift evenly distributed patches from 3 frames instead of uneven left half (ablation)PST patch-only · Something-Something V1top-1: 45.3 -> 48.6Improved
use pattern C-9 (9-frame temporal field) instead of A-3 (ablation)PST patch-only · Something-Something V1top-1: 48.6 -> 51.8Improved
use pattern D-16 instead of C-9 (ablation)PST patch-only · Something-Something V1top-1: 51.8 -> 50.0Hurt
add TPS to all four stages instead of stage 1 only (ablation)PST-T · Something-Something V1top-1: 47.3 -> 51.8Improved
add shift back (ablation)PST-T · Something-Something V1top-1: 47.3 -> 51.8Improved
add alternative shift (ablation)PST-T · Something-Something V1top-1: 46.4 -> 51.8Improved
add shift of relative position encoding (ablation)PST-T · Something-Something V1top-1: 46.1 -> 51.8Improved
add channel shift to patch shift (ablation)PST-T · Something-Something V1top-1: 51.8 -> 52.2Improved
use PST instead of joint spatiotemporal SASwin-T · Something-Something V1top-1 / memory: 51.5 -> 52.2 top-1; 106G -> 72G FLOPs; 20.2G -> 3.7G memoryImproved
use TPS instead of average pooling temporal modelingDeiT-S · Kinetics-400top-1: 73.0 -> 75.3Improved
use TPS instead of temporal channel shift (TSM)DeiT-S · Kinetics-400top-1: 74.8 -> 75.3Improved

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