COMPOSER: Compositional Reasoning of Group Activity in Videos with Keypoint-Only Modality

Honglu Zhou, Asim Kadav, Aviv Shamsian, Shijie Geng, Farley Lai, Long Zhao, Ting Liu, Mubbasir Kapadia, Hans Peter Graf

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

5 Scopus citations

Abstract

Group Activity Recognition detects the activity collectively performed by a group of actors, which requires compositional reasoning of actors and objects. We approach the task by modeling the video as tokens that represent the multi-scale semantic concepts in the video. We propose COMPOSER, a Multiscale Transformer based architecture that performs attention-based reasoning over tokens at each scale and learns group activity compositionally. In addition, prior works suffer from scene biases with privacy and ethical concerns. We only use the keypoint modality which reduces scene biases and prevents acquiring detailed visual data that may contain private or biased information of users. We improve the multiscale representations in COMPOSER by clustering the intermediate scale representations, while maintaining consistent cluster assignments between scales. Finally, we use techniques such as auxiliary prediction and data augmentations tailored to the keypoint signals to aid model training. We demonstrate the model’s strength and interpretability on two widely-used datasets (Volleyball and Collective Activity). COMPOSER achieves up to + 5.4 % improvement with just the keypoint modality (Code is available at https://github.com/hongluzhou/composer.).

Original languageEnglish
Title of host publicationComputer Vision – ECCV 2022 - 17th European Conference, Proceedings
EditorsShai Avidan, Gabriel Brostow, Moustapha Cissé, Giovanni Maria Farinella, Tal Hassner
PublisherSpringer Science and Business Media Deutschland GmbH
Pages249-266
Number of pages18
ISBN (Print)9783031198328
DOIs
StatePublished - 2022
Event17th European Conference on Computer Vision, ECCV 2022 - Tel Aviv, Israel
Duration: 23 Oct 202227 Oct 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13695 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference17th European Conference on Computer Vision, ECCV 2022
Country/TerritoryIsrael
CityTel Aviv
Period23/10/2227/10/22

Bibliographical note

Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Keywords

  • Compositionality
  • Keypoint-only group activity recognition
  • Multiscale representations
  • Transformer
  • Video understanding

Fingerprint

Dive into the research topics of 'COMPOSER: Compositional Reasoning of Group Activity in Videos with Keypoint-Only Modality'. Together they form a unique fingerprint.

Cite this