Flow-Assisted Motion Learning Network for Weakly-Supervised Group Activity Recognition

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Nugroho, Muhammad Adi, Woo, Sangmin, Lee, Sumin, Park, Jinyoung, Wang, Yooseung, Kim, Donguk, Kim, Changick
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914814749573120
author Nugroho, Muhammad Adi
Woo, Sangmin
Lee, Sumin
Park, Jinyoung
Wang, Yooseung
Kim, Donguk
Kim, Changick
author_facet Nugroho, Muhammad Adi
Woo, Sangmin
Lee, Sumin
Park, Jinyoung
Wang, Yooseung
Kim, Donguk
Kim, Changick
contents Weakly-Supervised Group Activity Recognition (WSGAR) aims to understand the activity performed together by a group of individuals with the video-level label and without actor-level labels. We propose Flow-Assisted Motion Learning Network (Flaming-Net) for WSGAR, which consists of the motion-aware actor encoder to extract actor features and the two-pathways relation module to infer the interaction among actors and their activity. Flaming-Net leverages an additional optical flow modality in the training stage to enhance its motion awareness when finding locally active actors. The first pathway of the relation module, the actor-centric path, initially captures the temporal dynamics of individual actors and then constructs inter-actor relationships. In parallel, the group-centric path starts by building spatial connections between actors within the same timeframe and then captures simultaneous spatio-temporal dynamics among them. We demonstrate that Flaming-Net achieves new state-of-the-art WSGAR results on two benchmarks, including a 2.8%p higher MPCA score on the NBA dataset. Importantly, we use the optical flow modality only for training and not for inference.
format Preprint
id arxiv_https___arxiv_org_abs_2405_18012
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Flow-Assisted Motion Learning Network for Weakly-Supervised Group Activity Recognition
Nugroho, Muhammad Adi
Woo, Sangmin
Lee, Sumin
Park, Jinyoung
Wang, Yooseung
Kim, Donguk
Kim, Changick
Computer Vision and Pattern Recognition
Image and Video Processing
Weakly-Supervised Group Activity Recognition (WSGAR) aims to understand the activity performed together by a group of individuals with the video-level label and without actor-level labels. We propose Flow-Assisted Motion Learning Network (Flaming-Net) for WSGAR, which consists of the motion-aware actor encoder to extract actor features and the two-pathways relation module to infer the interaction among actors and their activity. Flaming-Net leverages an additional optical flow modality in the training stage to enhance its motion awareness when finding locally active actors. The first pathway of the relation module, the actor-centric path, initially captures the temporal dynamics of individual actors and then constructs inter-actor relationships. In parallel, the group-centric path starts by building spatial connections between actors within the same timeframe and then captures simultaneous spatio-temporal dynamics among them. We demonstrate that Flaming-Net achieves new state-of-the-art WSGAR results on two benchmarks, including a 2.8%p higher MPCA score on the NBA dataset. Importantly, we use the optical flow modality only for training and not for inference.
title Flow-Assisted Motion Learning Network for Weakly-Supervised Group Activity Recognition
topic Computer Vision and Pattern Recognition
Image and Video Processing
url https://arxiv.org/abs/2405.18012