</> Code with Papers by Syntology

Papers › Temporally Consistent Unbalanced Optimal Transport for Unsupervised Action Segmentation

Temporally Consistent Unbalanced Optimal Transport for Unsupervised Action Segmentation

1 Apr 2024CVPR 2024 1arXiv:2404.01518archive 2025-07-28

Ming Xu, Stephen Gould

We propose a novel approach to the action segmentation task for long, untrimmed videos, based on solving an optimal transport problem. By encoding a temporal consistency prior into a Gromov-Wasserstein problem, we are able to decode a temporally consistent segmentation from a noisy affinity/matching cost matrix between video frames and action classes. Unlike previous approaches, our method does not require knowing the action order for a video to attain temporal consistency. Furthermore, our resulting (fused) Gromov-Wasserstein problem can be efficiently solved on GPUs using a few iterations of projected mirror descent. We demonstrate the effectiveness of our method in an unsupervised learning setting, where our method is used to generate pseudo-labels for self-training. We evaluate our segmentation approach and unsupervised learning pipeline on the Breakfast, 50-Salads, YouTube Instructions and Desktop Assembly datasets, yielding state-of-the-art results for the unsupervised video action segmentation task.

PaperPDFConference PDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

mingu6/action_seg_ot officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Action SegmentationSegmentationUnsupervised Action Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Unsupervised Action Segmentation Breakfast ASOT Acc 56.1 #2 of 8 Archive leaderboard report
Unsupervised Action Segmentation Breakfast ASOT F1 38.3 #2 of 8 Archive leaderboard report
Unsupervised Action Segmentation Breakfast ASOT JSD 94.9 #2 of 8 Archive leaderboard report
Unsupervised Action Segmentation Breakfast ASOT Precision 36.7 #2 of 8 Archive leaderboard report
Unsupervised Action Segmentation Breakfast ASOT Recall 40.1 #2 of 8 Archive leaderboard report
Unsupervised Action Segmentation Breakfast ASOT mIoU 18.6 #2 of 8 Archive leaderboard report
Unsupervised Action Segmentation IKEA ASM ASOT Accuracy 34.0 #2 of 5 Archive leaderboard report
Unsupervised Action Segmentation IKEA ASM ASOT F1 27.9 #2 of 5 Archive leaderboard report
Unsupervised Action Segmentation IKEA ASM ASOT JSD 88.7 #2 of 5 Archive leaderboard report
Unsupervised Action Segmentation IKEA ASM ASOT Precision 21.1 #2 of 5 Archive leaderboard report
Unsupervised Action Segmentation IKEA ASM ASOT Recall 24.0 #2 of 5 Archive leaderboard report
Unsupervised Action Segmentation Youtube INRIA Instructional ASOT Acc 52.9 #2 of 8 Archive leaderboard report
Unsupervised Action Segmentation Youtube INRIA Instructional ASOT F1 35.1 #2 of 8 Archive leaderboard report
Unsupervised Action Segmentation Youtube INRIA Instructional ASOT Precision 47.6 #2 of 8 Archive leaderboard report
Unsupervised Action Segmentation Youtube INRIA Instructional ASOT Recall 27.8 #2 of 8 Archive leaderboard report
Unsupervised Action Segmentation Youtube INRIA Instructional ASOT mIoU 24.7 #2 of 8 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections