Invariant Classification of Gait Types

Preben Fihl, Thomas B. Moeslund

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

327 Downloads (Pure)

Abstract

This paper presents a method of classifying human gait in an invariant manner based on silhouette comparison. A database of artificially generated silhouettes is created representing the three main types of gait, i.e. walking, jogging, and running. Silhouettes generated from different camera angles are included in the database to make the method invariant to camera viewpoint and to changing directions of movement. The extraction of silhouettes are done using the Codebook method and silhouettes are represented in a scale- and translation-invariant manner by using shape contexts and tangent orientations. Input silhouettes are matched to the database using the Hungarian method. A classifier is defined based on the dissimilarity between the input silhouettes and the gait actions of the database. The overall recognition rate is 88.2% on a large and diverse test set. The recognition rate is better than that achieved by other approaches applied to similar data.

Original languageEnglish
Title of host publicationCanadian Conference on Computer and Robot Vision, 2008. CRV '08
Number of pages7
PublisherIEEE
Publication date2008
Pages179-185
ISBN (Print)978-0-7695-3153-3
DOIs
Publication statusPublished - 2008
EventCanadian Conference on Computer and Robot Vision - Windsor, Ontario, Canada
Duration: 28 May 200830 May 2008
Conference number: 5th

Conference

ConferenceCanadian Conference on Computer and Robot Vision
Number5th
CountryCanada
CityWindsor, Ontario
Period28/05/200830/05/2008

Fingerprint

Cameras
Classifiers

Keywords

  • Action Recognition
  • Human motion
  • Gait Analysis

Cite this

Fihl, P., & Moeslund, T. B. (2008). Invariant Classification of Gait Types. In Canadian Conference on Computer and Robot Vision, 2008. CRV '08 (pp. 179-185). IEEE. https://doi.org/10.1109/CRV.2008.24
Fihl, Preben ; Moeslund, Thomas B. / Invariant Classification of Gait Types. Canadian Conference on Computer and Robot Vision, 2008. CRV '08. IEEE, 2008. pp. 179-185
@inproceedings{fae26130993511deb96d000ea68e967b,
title = "Invariant Classification of Gait Types",
abstract = "This paper presents a method of classifying human gait in an invariant manner based on silhouette comparison. A database of artificially generated silhouettes is created representing the three main types of gait, i.e. walking, jogging, and running. Silhouettes generated from different camera angles are included in the database to make the method invariant to camera viewpoint and to changing directions of movement. The extraction of silhouettes are done using the Codebook method and silhouettes are represented in a scale- and translation-invariant manner by using shape contexts and tangent orientations. Input silhouettes are matched to the database using the Hungarian method. A classifier is defined based on the dissimilarity between the input silhouettes and the gait actions of the database. The overall recognition rate is 88.2{\%} on a large and diverse test set. The recognition rate is better than that achieved by other approaches applied to similar data.",
keywords = "Action Recognition, Human motion, Gait Analysis",
author = "Preben Fihl and Moeslund, {Thomas B.}",
year = "2008",
doi = "10.1109/CRV.2008.24",
language = "English",
isbn = "978-0-7695-3153-3",
pages = "179--185",
booktitle = "Canadian Conference on Computer and Robot Vision, 2008. CRV '08",
publisher = "IEEE",
address = "United States",

}

Fihl, P & Moeslund, TB 2008, Invariant Classification of Gait Types. in Canadian Conference on Computer and Robot Vision, 2008. CRV '08. IEEE, pp. 179-185, Windsor, Ontario, Canada, 28/05/2008. https://doi.org/10.1109/CRV.2008.24

Invariant Classification of Gait Types. / Fihl, Preben; Moeslund, Thomas B.

Canadian Conference on Computer and Robot Vision, 2008. CRV '08. IEEE, 2008. p. 179-185.

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

TY - GEN

T1 - Invariant Classification of Gait Types

AU - Fihl, Preben

AU - Moeslund, Thomas B.

PY - 2008

Y1 - 2008

N2 - This paper presents a method of classifying human gait in an invariant manner based on silhouette comparison. A database of artificially generated silhouettes is created representing the three main types of gait, i.e. walking, jogging, and running. Silhouettes generated from different camera angles are included in the database to make the method invariant to camera viewpoint and to changing directions of movement. The extraction of silhouettes are done using the Codebook method and silhouettes are represented in a scale- and translation-invariant manner by using shape contexts and tangent orientations. Input silhouettes are matched to the database using the Hungarian method. A classifier is defined based on the dissimilarity between the input silhouettes and the gait actions of the database. The overall recognition rate is 88.2% on a large and diverse test set. The recognition rate is better than that achieved by other approaches applied to similar data.

AB - This paper presents a method of classifying human gait in an invariant manner based on silhouette comparison. A database of artificially generated silhouettes is created representing the three main types of gait, i.e. walking, jogging, and running. Silhouettes generated from different camera angles are included in the database to make the method invariant to camera viewpoint and to changing directions of movement. The extraction of silhouettes are done using the Codebook method and silhouettes are represented in a scale- and translation-invariant manner by using shape contexts and tangent orientations. Input silhouettes are matched to the database using the Hungarian method. A classifier is defined based on the dissimilarity between the input silhouettes and the gait actions of the database. The overall recognition rate is 88.2% on a large and diverse test set. The recognition rate is better than that achieved by other approaches applied to similar data.

KW - Action Recognition

KW - Human motion

KW - Gait Analysis

U2 - 10.1109/CRV.2008.24

DO - 10.1109/CRV.2008.24

M3 - Article in proceeding

SN - 978-0-7695-3153-3

SP - 179

EP - 185

BT - Canadian Conference on Computer and Robot Vision, 2008. CRV '08

PB - IEEE

ER -

Fihl P, Moeslund TB. Invariant Classification of Gait Types. In Canadian Conference on Computer and Robot Vision, 2008. CRV '08. IEEE. 2008. p. 179-185 https://doi.org/10.1109/CRV.2008.24