Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Exploring High-Order Self-Similarity for Video Understanding

About

Space-time self-similarity (STSS), which captures visual correspondences across frames, provides an effective way to represent temporal dynamics for video understanding. In this work, we explore higher-order STSS and demonstrate how STSSs at different orders reveal distinct aspects of these dynamics. We then introduce the Multi-Order Self-Similarity (MOSS) module, a lightweight neural module designed to learn and integrate multi-order STSS features. It can be applied to diverse video tasks to enhance motion modeling capabilities while consuming only marginal computational cost and memory usage. Extensive experiments on video action recognition, motion-centric video VQA, and real-world robotic tasks consistently demonstrate substantial improvements, validating the broad applicability of MOSS as a general temporal modeling module. The source code and checkpoints will be publicly available.

Manjin Kim, Heeseung Kwon, Karteek Alahari, Minsu Cho• 2026

Related benchmarks

TaskDatasetResultRank
Action RecognitionKinetics-400
Top-1 Acc87.7
505
Temporal Action DetectionTHUMOS-14 (test)
mAP@tIoU=0.577.6
339
Action RecognitionDiving-48 (test)
Top-1 Acc92.7
92
Video ClassificationSomething-Something v2
Top-1 Acc75.3
78
Video Action ClassificationDiving-48
Top-1 Acc92.7
64
Video UnderstandingFAVOR-Bench
AS46.8
37
Action RecognitionFineGym Gym288--
24
Video ClassificationSomething-Something V1
Top-1 Acc67.3
23
Action RecognitionGym99
Top-1 Accuracy94.7
19
Generic Event Boundary DetectionKinetics-GEBD
F1@0.0579.4
17
Showing 10 of 15 rows

Other info

Follow for update