AI Paper Atlas

Research map / Video / Action recognition

Two-Stage Active Learning for Efficient Temporal Action Segmentation

Yuhao Su, Ehsan Elhamifar

ECCV 2024 · Video › Action recognition · approach: Temporal action segmentation · 8 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 regularizers L_mu and L_sigma to contrastive prototype loss (ablation)Action prototype learning in TAS · 50Saladsaccuracy: 47.2 -> 57.8Improved
add regularizers L_mu and L_sigma to contrastive prototype loss (ablation)Action prototype learning in TAS · GTEAaccuracy: 43.9 -> 47.6Improved
use L_cont alone instead of plain average-embedding prototypes (ablation)Action prototype learning in TAS · 50Saladsaccuracy: 50.0 -> 47.2Hurt
use VAS intra-video selection instead of split-random (ablation)Two-stage AL with VPA · 50Saladsaccuracy: 45.4 -> 57.8Improved
use VPA inter-video selection instead of random selection (ablation)Two-stage AL with VAS · 50Saladsaccuracy: 44.7 -> 57.8Improved
use proposed two-stage selection versus split-random baselineTAS active learning · 50Salads (0.16% frames)f1@50: 24.7 -> 32.9Improved
use proposed selection on CrossTask with 72% background framesTAS active learning · CrossTask (0.16% frames)f1@50: 11.2 -> 12.4Improved
add pseudo-labels for low-alignment-cost videos from 5th iterationTwo-stage AL on Breakfast · Breakfast (0.31% frames)f1@50 as percent of full supervision: 90% -> close to 95%Improved
keep adding pseudo-labels beyond 0.35% budgetTwo-stage AL on Breakfast · Breakfast (>0.35% frames)f1@50Hurt
use temporally separated frame labeling baselines instead of coresetIntra-video selection baselines · all four datasetsTAS performance: 38.4 -> 49.0 acc (50Salads, coreset vs split-rand)Improved
label 0.23% frames with our method versus all-video timestamp supervision (0.29%)Active learning vs timestamp supervision · Breakfastf1@50: 50.7 -> 54.3 (D-TSTAS vs ours)Improved

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