AI Paper Atlas

Research map / Video / Action recognition

VidTr: Video Transformer Without Convolutions

Yanyi Zhang, Xinyu Li, Chunhui Liu, Bing Shuai, Yi Zhu, Biagio Brattoli, Hao Chen, Ivan Marsic, Joseph Tighe

ICCV 2021 · Video › Action recognition · approach: Spatiotemporal CNNs for action recognition · 231 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
use square 1x16^2 patches instead of cubic 4x16^2 patches (ablation)VidTr-S · Kinetics-400top-1 accuracy: 73.1 -> 77.7Improved
use 16x16 patches instead of 32x32 patches (ablation)VidTr-S · Kinetics-400top-1 accuracy: 71.2 -> 77.7Improved
add temporal attention to spatial-only transformer (WH to WH + T) (ablation)VidTr-S · Kinetics-400top-1 accuracy: 74.7 -> 77.7Improved
separable attention (WH + T) instead of joint spatio-temporal attention (WHT) (ablation)VidTr-S · Kinetics-400memory / top-1 accuracy: 7.6GB -> 2.3GB memory, 77.5 -> 77.7Improved
axial attention (W + H + T) instead of WH + T (ablation)VidTr-S · Kinetics-400top-1 accuracy: 77.7 -> 72.3Hurt
STD topK pooling instead of temporal average pooling (ablation)C-VidTr-S · Kinetics-400top-1 accuracy: 74.9 -> 75.7Improved
STD topK pooling instead of 1D strided convolution (ablation)C-VidTr-S · Kinetics-400top-1 accuracy: 75.4 -> 75.7Improved
start temporal down-sampling at layer 1 instead of layer 0 (ablation)C-VidTr-S · Kinetics-400top-1 accuracy: 72.9 -> 74.9Improved
skip one layer between down-sample layers instead of none (ablation)C-VidTr-S · Kinetics-400top-1 accuracy: 73.9 -> 74.9Improved
skip more layers between down-sample layers (down-sample later) (ablation)C-VidTr-S · Kinetics-400top-1 accuracy / FLOPs: 74.9 -> 75.2No clear effect
initialize from ViT-L instead of ViT-B (ablation)VidTr-S · Kinetics-400top-1 accuracy: 77.7 -> 77.5No clear effect
ensemble VidTr with I3D50VidTr-S · Kinetics-400top-1 accuracy: 77.7 -> 79.4Improved

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