Skip to yearly menu bar Skip to main content


Temporally Consistent Unbalanced Optimal Transport for Unsupervised Action Segmentation

Ming Xu · Stephen Gould

Arch 4A-E Poster #400
[ ] [ Project Page ]
Thu 20 Jun 5 p.m. PDT — 6:30 p.m. PDT
Oral presentation: Orals 4C Action and motion
Thu 20 Jun 1 p.m. PDT — 2:30 p.m. PDT


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.

Live content is unavailable. Log in and register to view live content