Human Group Activity Recognition based on Modelling Moving Regions Interdependencies

Kyle Stephens; Adrian Gheorghe Bors

doi:10.1109/ICPR.2016.7899948

Human Group Activity Recognition based on Modelling Moving Regions Interdependencies

Kyle Stephens, Adrian Gheorghe Bors

Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

n this research study, we model the interdepen-
dency of actions performed by people in a group in order to
identify their activity. Unlike single human activity recognition,
in interacting groups the local movement activity is usually
influenced by the other persons in the group. We propose a
model to describe the discriminative characteristics of group
activity by considering the relations between motion flows and
the locations of moving regions. The inputs of the proposed model
are jointly represented in time-space and time-movement spaces.
These spaces are modelled using Kernel Density Estimation
(KDE) which is then fed into a machine learning classifier. Unlike
in other group-based human activity recognition algorithms, the
proposed methodology is automatic and does not rely on any
pedestrian detection or on the manual annotation of tracks.
Index Terms
—Group Activity Identification, Motion Segme

Original language	English
Title of host publication	Proc. of International Conference on Pattern Recognition (ICPR)
Publisher	IEEE
Pages	2116-2121
Number of pages	6
DOIs	https://doi.org/10.1109/ICPR.2016.7899948
Publication status	Published - 27 Apr 2017

Bibliographical note

© Copyright 2016 IEEE - All rights reserved. Use of this web site signifies your agreement to the terms and conditions. This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s

Access to Document

10.1109/ICPR.2016.7899948

Human Group Activity Recognition based on Modelling Moving Regions Interdependencies
© Copyright 2016 IEEE - All rights reserved. Use of this web site signifies your agreement to the terms and conditions. This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s
Accepted author manuscript, 558 KB

Cite this

@inproceedings{6f13e2b49c2541d4bd4cb5c79cf9b82d,

title = "Human Group Activity Recognition based on Modelling Moving Regions Interdependencies",

abstract = "n this research study, we model the interdepen-dency of actions performed by people in a group in order toidentify their activity. Unlike single human activity recognition,in interacting groups the local movement activity is usuallyinfluenced by the other persons in the group. We propose amodel to describe the discriminative characteristics of groupactivity by considering the relations between motion flows andthe locations of moving regions. The inputs of the proposed modelare jointly represented in time-space and time-movement spaces.These spaces are modelled using Kernel Density Estimation(KDE) which is then fed into a machine learning classifier. Unlikein other group-based human activity recognition algorithms, theproposed methodology is automatic and does not rely on anypedestrian detection or on the manual annotation of tracks.Index Terms—Group Activity Identification, Motion Segme",

author = "Kyle Stephens and Bors, {Adrian Gheorghe}",

note = "{\textcopyright} Copyright 2016 IEEE - All rights reserved. Use of this web site signifies your agreement to the terms and conditions. This is an author-produced version of the published paper. Uploaded in accordance with the publisher{\textquoteright}s ",

year = "2017",

month = apr,

day = "27",

doi = "10.1109/ICPR.2016.7899948",

language = "English",

pages = "2116--2121",

booktitle = "Proc. of International Conference on Pattern Recognition (ICPR)",

publisher = "IEEE",

address = "United States",

}

TY - GEN

T1 - Human Group Activity Recognition based on Modelling Moving Regions Interdependencies

AU - Stephens, Kyle

AU - Bors, Adrian Gheorghe

N1 - © Copyright 2016 IEEE - All rights reserved. Use of this web site signifies your agreement to the terms and conditions. This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s

PY - 2017/4/27

Y1 - 2017/4/27

N2 - n this research study, we model the interdepen-dency of actions performed by people in a group in order toidentify their activity. Unlike single human activity recognition,in interacting groups the local movement activity is usuallyinfluenced by the other persons in the group. We propose amodel to describe the discriminative characteristics of groupactivity by considering the relations between motion flows andthe locations of moving regions. The inputs of the proposed modelare jointly represented in time-space and time-movement spaces.These spaces are modelled using Kernel Density Estimation(KDE) which is then fed into a machine learning classifier. Unlikein other group-based human activity recognition algorithms, theproposed methodology is automatic and does not rely on anypedestrian detection or on the manual annotation of tracks.Index Terms—Group Activity Identification, Motion Segme

AB - n this research study, we model the interdepen-dency of actions performed by people in a group in order toidentify their activity. Unlike single human activity recognition,in interacting groups the local movement activity is usuallyinfluenced by the other persons in the group. We propose amodel to describe the discriminative characteristics of groupactivity by considering the relations between motion flows andthe locations of moving regions. The inputs of the proposed modelare jointly represented in time-space and time-movement spaces.These spaces are modelled using Kernel Density Estimation(KDE) which is then fed into a machine learning classifier. Unlikein other group-based human activity recognition algorithms, theproposed methodology is automatic and does not rely on anypedestrian detection or on the manual annotation of tracks.Index Terms—Group Activity Identification, Motion Segme

U2 - 10.1109/ICPR.2016.7899948

DO - 10.1109/ICPR.2016.7899948

M3 - Conference contribution

SP - 2116

EP - 2121

BT - Proc. of International Conference on Pattern Recognition (ICPR)

PB - IEEE

ER -