Papers › Hierarchical Vector Quantization for Unsupervised Action Segmentation

Hierarchical Vector Quantization for Unsupervised Action Segmentation

23 Dec 2024arXiv:2412.17640archive 2025-07-28

Federico Spurio, Emad Bahrami, Gianpiero Francesca, Juergen Gall

In this work, we address unsupervised temporal action segmentation, which segments a set of long, untrimmed videos into semantically meaningful segments that are consistent across videos. While recent approaches combine representation learning and clustering in a single step for this task, they do not cope with large variations within temporal segments of the same class. To address this limitation, we propose a novel method, termed Hierarchical Vector Quantization (HVQ), that consists of two subsequent vector quantization modules. This results in a hierarchical clustering where the additional subclusters cover the variations within a cluster. We demonstrate that our approach captures the distribution of segment lengths much better than the state of the art. To this end, we introduce a new metric based on the Jensen-Shannon Distance (JSD) for unsupervised temporal action segmentation. We evaluate our approach on three public datasets, namely Breakfast, YouTube Instructional and IKEA ASM. Our approach outperforms the state of the art in terms of F1 score, recall and JSD.

PaperPDFCode

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

Code

fedespu/hvq 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 SegmentationClusteringQuantizationRepresentation LearningTemporal Action SegmentationUnsupervised Action Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Unsupervised Action Segmentation Breakfast HVQ Acc 54.4 #1 of 8 Archive leaderboard report
Unsupervised Action Segmentation Breakfast HVQ F1 39.7 #1 of 8 Archive leaderboard report
Unsupervised Action Segmentation Breakfast HVQ JSD 82.5 #1 of 8 Archive leaderboard report
Unsupervised Action Segmentation Breakfast HVQ Precision 35.6 #1 of 8 Archive leaderboard report
Unsupervised Action Segmentation Breakfast HVQ Recall 44.9 #1 of 8 Archive leaderboard report
Unsupervised Action Segmentation IKEA ASM HVQ Accuracy 51.2 #1 of 5 Archive leaderboard report
Unsupervised Action Segmentation IKEA ASM HVQ F1 30.7 #1 of 5 Archive leaderboard report
Unsupervised Action Segmentation IKEA ASM HVQ JSD 64.8 #1 of 5 Archive leaderboard report
Unsupervised Action Segmentation IKEA ASM HVQ Precision 37.7 #1 of 5 Archive leaderboard report
Unsupervised Action Segmentation IKEA ASM HVQ Recall 25.9 #1 of 5 Archive leaderboard report
Unsupervised Action Segmentation Youtube INRIA Instructional HVQ Acc 50.3 #3 of 8 Archive leaderboard report
Unsupervised Action Segmentation Youtube INRIA Instructional HVQ F1 35.1 #3 of 8 Archive leaderboard report
Unsupervised Action Segmentation Youtube INRIA Instructional HVQ Precision 32.1 #3 of 8 Archive leaderboard report
Unsupervised Action Segmentation Youtube INRIA Instructional HVQ Recall 38.7 #3 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