AI Paper Atlas

Research map / Video / Action recognition

Uncertainty-Aware Weakly Supervised Action Detection from Untrimmed Videos

Anurag Arnab, Chen Sun, Arsha Nagrani, Cordelia Schmid

ECCV 2020 · Video › Action recognition · approach: Temporal action segmentation · 31 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 multiple instance learning (max pooling) over tubelets instead of using video label per tubelet (ablation)weakly supervised action detector · UCF101-24 valVideo AP@0.5: 29.7 -> 33.5Improved
add uncertainty-based loss to MIL max pooling (ablation)MIL max pooling · UCF101-24 valVideo AP@0.5: 33.5 -> 35.0Improved
add uncertainty-based loss to MIL max pooling (ablation)MIL max pooling · UCF101-24 valVideo AP@0.2: 60.7 -> 61.7Improved
use max pooling instead of mean pooling (ablation)MIL · UCF101-24 valVideo AP@0.5: 33.0 -> 33.5Improved
use max pooling instead of LSE pooling (ablation)MIL · UCF101-24 valVideo AP@0.5: 33.1 -> 33.5Improved
aggregation function choice (LSE, mean, max) (ablation)MIL · UCF101-24 valVideo AP@0.2: 60.1 / 60.3 / 60.7No clear effect
add uncertainty loss with 3 bags x 5 tubelets (ablation)MIL max pooling · UCF101-24 valVideo AP@0.5: 33.6 -> 34.1Improved
add uncertainty loss with 2 bags x 8 tubelets (ablation)MIL max pooling · UCF101-24 valVideo AP@0.5: 33.3 -> 34.2Improved
add uncertainty loss with 1 bag x 16 tubelets (ablation)MIL max pooling · UCF101-24 valVideo AP@0.5: 25.8 -> 26.2Improved
use fewer, larger bags (1 bag of 16 tubelets instead of 4 bags of 4) (ablation)MIL max pooling + uncertainty · UCF101-24 valVideo AP@0.5: 35.0 -> 26.2Hurt
shorten the sub-clip duration of weak labels from 900 s to 1 sMIL max pooling + uncertainty · AVA v2.2Frame AP@0.5: 4.2 -> 22.4Improved
full box supervision instead of weak 5 s sub-clip labelsSlowFast R50 · AVA v2.2Frame AP@0.5: 18.0 -> 24.9Improved

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