AI Paper Atlas

Research map / Video / Action recognition

Fast Weakly Supervised Action Segmentation Using Mutual Consistency

Yaser Souri, Mohsen Fayyaz, Luca Minciullo, Gianpiero Francesca, Jüergen Gall

TPAMI 2021 · Video › Action recognition · approach: Temporal action segmentation

Read the paper

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 frame branch and mutual consistency loss (ablation)segment generation branch · Breakfast split 1matching score: 0.729 -> 0.774Improved
use exponentially increasing dilation in temporal backbone instead of single conv (ablation)segment branch · Breakfast split 1matching score: 0.691 -> 0.729Improved
use box mask instead of bell mask (ablation)MuCon · Breakfast split 1MoF accuracy: 43.6 -> 49.0Improved
add 10% overlap to box masks (ablation)MuCon · Breakfast split 1MoF accuracy: 49.0 -> 48.8No clear effect
add length regularizer and smoothing loss (ablation)MuCon · Breakfast split 1MoF accuracy: 47.4 -> 49.0Improved
add smoothing loss only (ablation)MuCon · Breakfast split 1MoF accuracy: 47.4 -> 47.5No clear effect
fuse both branches at inference (MuCon-full) instead of segment branch only (ablation)MuCon · Breakfast split 1MoF accuracy: 43.6 -> 49.0Improved
add teacher forcing in training (ablation)MuCon · Breakfast split 1MoF accuracy: 48.7 -> 49.0No clear effect
add MuCon loss in fully supervised training (ablation)MuCon · Breakfast split 1 (fully supervised)MoF accuracy: 63.7 -> 64.1Improved
add 10% fully annotated videos to weak supervision (ablation)MuCon · Breakfast split 1MoF accuracy: 49.0 -> 56.2Improved
add 1% fully annotated videos to weak supervision (ablation)MuCon · Breakfast split 1MoF accuracy: 49.0 -> 51.0No clear effect
use MuCon instead of CDFLweakly supervised action segmentation · Breakfasttraining time (hours): 66.73 -> 4.57Improved

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