AI Paper Atlas

Research map / Video / Action recognition

How Much Temporal Long-Term Context is Needed for Action Segmentation?

Emad Bahrami, Gianpiero Francesca, Juergen Gall

ICCV 2023 · Video › Action recognition · approach: Fully supervised temporal action detection · 58 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
combine windowed and long-term context attention in each block (ablation)LTContext · 50SaladsF1@50: 77.4 -> 82.0 (windowed only); 74.1 -> 82.0 (LTC only)Improved
increase window size W from 8 to 64 (ablation)LTContext · 50SaladsF1@50: 74.5 -> 82.0Improved
decrease G (denser sparse attention) from 64 to 8 (ablation)LTContext · 50SaladsF1@50: 82.0 -> 79.9Hurt
add overlap (2 windows) for keys and values in windowed attention (ablation)LTContext · 50SaladsF1@50: 79.1 -> 82.0Improved
extend overlap to 3 windows (ablation)LTContext · 50SaladsF1@50: 82.0 -> 80.8No clear effect
apply windowed attention before long-term context attention (ablation)LTContext · 50SaladsF1@50: 80.4 -> 82.0Improved
add cross-attention using predictions in stages 2-4 (ablation)LTContext · 50SaladsF1@50: 70.4 -> 82.0Improved
add dilated 1D convolution in the block (ablation)LTContext · 50SaladsF1@50: 67.1 -> 82.0 (no conv); 67.7 -> 82.0 (no dilation)Improved
use multiple attention heads instead of one (ablation)LTContext · 50SaladsF1@50: 82.0 -> 80.5 (2 heads)No clear effect
increase number of stages from 1 to 4 (ablation)LTContext · 50SaladsF1@50: 49.9 -> 82.0Improved
increase number of stages from 4 to 5 (ablation)LTContext · 50SaladsF1@50: 82.0 -> 80.0Hurt
increase layers per stage from 9 to 10 (ablation)LTContext · mixed datasetsF1@50: 50Salads 82.0 -> 82.4; Breakfast 60.1 -> 59.7; Assembly101 22.6 -> 21.9Mixed

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