Home > Research > Publications & Outputs > Discriminative human action classification usin...

Associated organisational unit

View graph of relations

Discriminative human action classification using locality-constrained linear coding

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published

Standard

Discriminative human action classification using locality-constrained linear coding. / Rahmani, Hossein; Huynh, Du Q.; Mahmood, Arif et al.

In: Pattern Recognition Letters, Vol. 72, 01.03.2016, p. 62-71.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Rahmani, H, Huynh, DQ, Mahmood, A & Mian, A 2016, 'Discriminative human action classification using locality-constrained linear coding', Pattern Recognition Letters, vol. 72, pp. 62-71. https://doi.org/10.1016/j.patrec.2015.07.015

APA

Vancouver

Rahmani H, Huynh DQ, Mahmood A, Mian A. Discriminative human action classification using locality-constrained linear coding. Pattern Recognition Letters. 2016 Mar 1;72:62-71. doi: 10.1016/j.patrec.2015.07.015

Author

Rahmani, Hossein ; Huynh, Du Q. ; Mahmood, Arif et al. / Discriminative human action classification using locality-constrained linear coding. In: Pattern Recognition Letters. 2016 ; Vol. 72. pp. 62-71.

Bibtex

@article{3c3d5ca9ca5c4fdc9aa4fe80e1b08909,
title = "Discriminative human action classification using locality-constrained linear coding",
abstract = "We propose a Locality-constrained Linear Coding (LLC) based algorithm that captures discriminative information of human actions in spatio-temporal subsequences of videos. The input video is divided into equally spaced overlapping spatio-temporal subsequences. Each subsequence is further divided into blocks and then cells. The spatio-temporal information in each cell is represented by a Histogram of Oriented 3D Gradients (HOG3D). LLC is then used to encode each block. We show that LLC gives more stable and repetitive codes compared to the standard Sparse Coding. The final representation of a video sequence is obtained using logistic regression with ℓ2regularization and classification is performed by a linear SVM. The proposed algorithm is applicable to conventional and depth videos. Experimental comparison with ten state-of-the-art methods on three depth video and two conventional video databases shows that the proposed method consistently achieves the best performance.",
keywords = "Human action classification, Locality-constrained linear coding, SVM classifier, Sparse coding",
author = "Hossein Rahmani and Huynh, {Du Q.} and Arif Mahmood and Ajmal Mian",
year = "2016",
month = mar,
day = "1",
doi = "10.1016/j.patrec.2015.07.015",
language = "English",
volume = "72",
pages = "62--71",
journal = "Pattern Recognition Letters",
issn = "0167-8655",
publisher = "Elsevier Science B.V.",

}

RIS

TY - JOUR

T1 - Discriminative human action classification using locality-constrained linear coding

AU - Rahmani, Hossein

AU - Huynh, Du Q.

AU - Mahmood, Arif

AU - Mian, Ajmal

PY - 2016/3/1

Y1 - 2016/3/1

N2 - We propose a Locality-constrained Linear Coding (LLC) based algorithm that captures discriminative information of human actions in spatio-temporal subsequences of videos. The input video is divided into equally spaced overlapping spatio-temporal subsequences. Each subsequence is further divided into blocks and then cells. The spatio-temporal information in each cell is represented by a Histogram of Oriented 3D Gradients (HOG3D). LLC is then used to encode each block. We show that LLC gives more stable and repetitive codes compared to the standard Sparse Coding. The final representation of a video sequence is obtained using logistic regression with ℓ2regularization and classification is performed by a linear SVM. The proposed algorithm is applicable to conventional and depth videos. Experimental comparison with ten state-of-the-art methods on three depth video and two conventional video databases shows that the proposed method consistently achieves the best performance.

AB - We propose a Locality-constrained Linear Coding (LLC) based algorithm that captures discriminative information of human actions in spatio-temporal subsequences of videos. The input video is divided into equally spaced overlapping spatio-temporal subsequences. Each subsequence is further divided into blocks and then cells. The spatio-temporal information in each cell is represented by a Histogram of Oriented 3D Gradients (HOG3D). LLC is then used to encode each block. We show that LLC gives more stable and repetitive codes compared to the standard Sparse Coding. The final representation of a video sequence is obtained using logistic regression with ℓ2regularization and classification is performed by a linear SVM. The proposed algorithm is applicable to conventional and depth videos. Experimental comparison with ten state-of-the-art methods on three depth video and two conventional video databases shows that the proposed method consistently achieves the best performance.

KW - Human action classification

KW - Locality-constrained linear coding

KW - SVM classifier

KW - Sparse coding

U2 - 10.1016/j.patrec.2015.07.015

DO - 10.1016/j.patrec.2015.07.015

M3 - Journal article

VL - 72

SP - 62

EP - 71

JO - Pattern Recognition Letters

JF - Pattern Recognition Letters

SN - 0167-8655

ER -