Feature integration with random forests for real-time human activity recognition

Hirokatsu Kataoka, Kiyoshi Hashimoto, Yoshimitsu Aoki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper presents an approach for real-time human activity recognition. Three different kinds of features (flow, shape, and a keypoint-based feature) are applied in activity recognition. We use random forests for feature integration and activity classification. A forest is created at each feature that performs as a weak classifier. The international classification of functioning, disability and health (ICF) proposed by WHO is applied in order to set the novel definition in activity recognition. Experiments on human activity recognition using the proposed framework show - 99.2% (Weizmann action dataset), 95.5% (KTH human actions dataset), and 54.6% (UCF50 dataset) recognition accuracy with a real-time processing speed. The feature integration and activity-class definition allow us to accomplish high-accuracy recognition match for the state-of-the-art in real-time.

Original languageEnglish
Title of host publicationProceedings of SPIE - The International Society for Optical Engineering
PublisherSPIE
Volume9445
ISBN (Print)9781628415605
DOIs
Publication statusPublished - 2015
Event7th International Conference on Machine Vision, ICMV 2014 - Milan, Italy
Duration: 2014 Nov 192014 Nov 21

Other

Other7th International Conference on Machine Vision, ICMV 2014
CountryItaly
CityMilan
Period14/11/1914/11/21

Fingerprint

Activity Recognition
Random Forest
Real-time
Classifiers
Health
Disability
Processing
disabilities
High Accuracy
Experiments
Classifier
classifiers
health
Human
Experiment

Keywords

  • Activity Recognition
  • Feature Integration
  • Random Forests

ASJC Scopus subject areas

  • Applied Mathematics
  • Computer Science Applications
  • Electrical and Electronic Engineering
  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics

Cite this

Kataoka, H., Hashimoto, K., & Aoki, Y. (2015). Feature integration with random forests for real-time human activity recognition. In Proceedings of SPIE - The International Society for Optical Engineering (Vol. 9445). [944506] SPIE. https://doi.org/10.1117/12.2181201

Feature integration with random forests for real-time human activity recognition. / Kataoka, Hirokatsu; Hashimoto, Kiyoshi; Aoki, Yoshimitsu.

Proceedings of SPIE - The International Society for Optical Engineering. Vol. 9445 SPIE, 2015. 944506.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Kataoka, H, Hashimoto, K & Aoki, Y 2015, Feature integration with random forests for real-time human activity recognition. in Proceedings of SPIE - The International Society for Optical Engineering. vol. 9445, 944506, SPIE, 7th International Conference on Machine Vision, ICMV 2014, Milan, Italy, 14/11/19. https://doi.org/10.1117/12.2181201
Kataoka H, Hashimoto K, Aoki Y. Feature integration with random forests for real-time human activity recognition. In Proceedings of SPIE - The International Society for Optical Engineering. Vol. 9445. SPIE. 2015. 944506 https://doi.org/10.1117/12.2181201
Kataoka, Hirokatsu ; Hashimoto, Kiyoshi ; Aoki, Yoshimitsu. / Feature integration with random forests for real-time human activity recognition. Proceedings of SPIE - The International Society for Optical Engineering. Vol. 9445 SPIE, 2015.
@inproceedings{bfc6371091c0429fadf142d5ca37cb4d,
title = "Feature integration with random forests for real-time human activity recognition",
abstract = "This paper presents an approach for real-time human activity recognition. Three different kinds of features (flow, shape, and a keypoint-based feature) are applied in activity recognition. We use random forests for feature integration and activity classification. A forest is created at each feature that performs as a weak classifier. The international classification of functioning, disability and health (ICF) proposed by WHO is applied in order to set the novel definition in activity recognition. Experiments on human activity recognition using the proposed framework show - 99.2{\%} (Weizmann action dataset), 95.5{\%} (KTH human actions dataset), and 54.6{\%} (UCF50 dataset) recognition accuracy with a real-time processing speed. The feature integration and activity-class definition allow us to accomplish high-accuracy recognition match for the state-of-the-art in real-time.",
keywords = "Activity Recognition, Feature Integration, Random Forests",
author = "Hirokatsu Kataoka and Kiyoshi Hashimoto and Yoshimitsu Aoki",
year = "2015",
doi = "10.1117/12.2181201",
language = "English",
isbn = "9781628415605",
volume = "9445",
booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",
publisher = "SPIE",

}

TY - GEN

T1 - Feature integration with random forests for real-time human activity recognition

AU - Kataoka, Hirokatsu

AU - Hashimoto, Kiyoshi

AU - Aoki, Yoshimitsu

PY - 2015

Y1 - 2015

N2 - This paper presents an approach for real-time human activity recognition. Three different kinds of features (flow, shape, and a keypoint-based feature) are applied in activity recognition. We use random forests for feature integration and activity classification. A forest is created at each feature that performs as a weak classifier. The international classification of functioning, disability and health (ICF) proposed by WHO is applied in order to set the novel definition in activity recognition. Experiments on human activity recognition using the proposed framework show - 99.2% (Weizmann action dataset), 95.5% (KTH human actions dataset), and 54.6% (UCF50 dataset) recognition accuracy with a real-time processing speed. The feature integration and activity-class definition allow us to accomplish high-accuracy recognition match for the state-of-the-art in real-time.

AB - This paper presents an approach for real-time human activity recognition. Three different kinds of features (flow, shape, and a keypoint-based feature) are applied in activity recognition. We use random forests for feature integration and activity classification. A forest is created at each feature that performs as a weak classifier. The international classification of functioning, disability and health (ICF) proposed by WHO is applied in order to set the novel definition in activity recognition. Experiments on human activity recognition using the proposed framework show - 99.2% (Weizmann action dataset), 95.5% (KTH human actions dataset), and 54.6% (UCF50 dataset) recognition accuracy with a real-time processing speed. The feature integration and activity-class definition allow us to accomplish high-accuracy recognition match for the state-of-the-art in real-time.

KW - Activity Recognition

KW - Feature Integration

KW - Random Forests

UR - http://www.scopus.com/inward/record.url?scp=84924368596&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84924368596&partnerID=8YFLogxK

U2 - 10.1117/12.2181201

DO - 10.1117/12.2181201

M3 - Conference contribution

SN - 9781628415605

VL - 9445

BT - Proceedings of SPIE - The International Society for Optical Engineering

PB - SPIE

ER -