</> Code with Papers by Syntology

Papers › MS-TCN: Multi-Stage Temporal Convolutional Network for Action Segmentation

MS-TCN: Multi-Stage Temporal Convolutional Network for Action Segmentation

5 Mar 2019CVPR 2019 6arXiv:1903.01945archive 2025-07-28

Yazan Abu Farha, Juergen Gall

Temporally locating and classifying action segments in long untrimmed videos is of particular interest to many applications like surveillance and robotics. While traditional approaches follow a two-step pipeline, by generating frame-wise probabilities and then feeding them to high-level temporal models, recent approaches use temporal convolutions to directly classify the video frames. In this paper, we introduce a multi-stage architecture for the temporal action segmentation task. Each stage features a set of dilated temporal convolutions to generate an initial prediction that is refined by the next one. This architecture is trained using a combination of a classification loss and a proposed smoothing loss that penalizes over-segmentation errors. Extensive evaluation shows the effectiveness of the proposed model in capturing long-range dependencies and recognizing action segments. Our model achieves state-of-the-art results on three challenging datasets: 50Salads, Georgia Tech Egocentric Activities (GTEA), and the Breakfast dataset.

PaperPDFConference PDFCode

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

Code

yabufarha/ms-tcn officialmentioned in papermentioned on GitHubpytorchNOASSERTION report
MindSpore-paper-code-2/code3 mindspore 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 SegmentationSegmentationTemporal Action Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Action Segmentation 50 Salads MS-TCN Acc 80.7 #27 of 28 Archive leaderboard report
Action Segmentation 50 Salads MS-TCN Edit 67.9 #27 of 28 Archive leaderboard report
Action Segmentation 50 Salads MS-TCN F1@10% 76.3 #27 of 28 Archive leaderboard report
Action Segmentation 50 Salads MS-TCN F1@25% 74.0 #27 of 28 Archive leaderboard report
Action Segmentation 50 Salads MS-TCN F1@50% 64.5 #27 of 28 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (IDT) Acc 65.1 #30 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (IDT) Average F1 50.6 #30 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (IDT) Edit 61.4 #30 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (IDT) F1@10% 58.2 #30 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (IDT) F1@25% 52.9 #30 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (IDT) F1@50% 40.8 #30 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (I3D) Acc 66.3 #31 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (I3D) Average F1 46.2 #31 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (I3D) Edit 61.7 #31 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (I3D) F1@10% 52.6 #31 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (I3D) F1@25% 48.1 #31 of 37 Archive leaderboard report
Action Segmentation Breakfast MS-TCN (I3D) F1@50% 37.9 #31 of 37 Archive leaderboard report
Action Segmentation GTEA MS-TCN Acc 79.2 #22 of 28 Archive leaderboard report
Action Segmentation GTEA MS-TCN Edit 81.4 #22 of 28 Archive leaderboard report
Action Segmentation GTEA MS-TCN F1@10% 87.5 #22 of 28 Archive leaderboard report
Action Segmentation GTEA MS-TCN F1@25% 85.4 #22 of 28 Archive leaderboard report
Action Segmentation GTEA MS-TCN F1@50% 74.6 #22 of 28 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