AI Paper Atlas

Research map / Video / Action recognition

Selective Feature Compression for Efficient Activity Recognition Inference

Chunhui Liu, Xinyu Li, Hao Chen, Davide Modolo, Joseph Tighe

ICCV 2021 · Video › Action recognition · approach: Disentangled action recognition · 10 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 30-crop dense sampling with SFCSlow-Only I3D-50 8x8 · Kinetics-400top-1 accuracy: 74.4 -> 74.6No clear effect
replace 30-crop dense sampling with SFCR(2+1)D-152 · Kinetics-400top-1 accuracy: 79.4 -> 80.0Improved
replace 30-crop dense sampling with SFCSlow-Only I3D-50 8x8 · Kinetics-400FLOPS: 1643G -> 247GImproved
use SFC with tau=2 instead of no compression (tau=1) (ablation)SFC I3D-50 8x8 · Kinetics-400 (I3D-50 8x8)top-1 accuracy: 73.8 -> 74.6Improved
compress more aggressively (tau=4 instead of tau=2) (ablation)SFC I3D-50 8x8 · Kinetics-400 (I3D-50 8x8)top-1 accuracy: 74.6 -> 72.0Hurt
insert SFC after Res3 instead of after Res1 (ablation)SFC I3D-50 8x8 · Kinetics-400 (I3D-50 8x8)top-1 accuracy: 73.1 -> 74.6Improved
use TopK pooling instead of average pooling (ablation)SFC I3D-50 8x8 · Kinetics-400 (I3D-50 8x8)top-1 accuracy: 72.9 -> 74.6Improved
use TopK pooling instead of max pooling (ablation)SFC I3D-50 8x8 · Kinetics-400 (I3D-50 8x8)top-1 accuracy: 72.6 -> 74.6Improved
add abstraction layer on q and k only (value untouched) (ablation)SFC I3D-50 8x8 · Kinetics-400 (I3D-50 8x8)top-1 accuracy: 72.7 -> 74.6Improved
also transform value v with the abstraction layer (ablation)SFC I3D-50 8x8 · Kinetics-400 (I3D-50 8x8)top-1 accuracy: 74.6 -> 63.2Hurt
use 3x1x1 temporal conv instead of 3x3x3 (ablation)SFC I3D-50 8x8 · Kinetics-400 (I3D-50 8x8)top-1 accuracy: 72.0 -> 74.6Improved
use 3x1x1 conv instead of 1x1x1 (ablation)SFC I3D-50 8x8 · Kinetics-400 (I3D-50 8x8)top-1 accuracy: 72.5 -> 74.6Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 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