Home > Research > Publications & Outputs > Human action recognition using deep rule-based ...

Electronic data

  • DRB_MTAP_D_19_02987_V0

    Rights statement: The final publication is available at Springer via http://dx.doi.org/10.1007/s11042-020-09381-9

    Accepted author manuscript, 1.76 MB, PDF document

    Available under license: CC BY-NC: Creative Commons Attribution-NonCommercial 4.0 International License

Links

Text available via DOI:

View graph of relations

Human action recognition using deep rule-based classifier

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published

Standard

Human action recognition using deep rule-based classifier. / Bux, Allah; Gu, Xiaowei; Angelov, Plamen et al.

In: Multimedia Tools and Applications, Vol. 79, No. 41-42, 01.11.2020, p. 30653-30667.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Bux, A, Gu, X, Angelov, P & Habib, Z 2020, 'Human action recognition using deep rule-based classifier', Multimedia Tools and Applications, vol. 79, no. 41-42, pp. 30653-30667. https://doi.org/10.1007/s11042-020-09381-9

APA

Vancouver

Bux A, Gu X, Angelov P, Habib Z. Human action recognition using deep rule-based classifier. Multimedia Tools and Applications. 2020 Nov 1;79(41-42):30653-30667. Epub 2020 Aug 17. doi: 10.1007/s11042-020-09381-9

Author

Bux, Allah ; Gu, Xiaowei ; Angelov, Plamen et al. / Human action recognition using deep rule-based classifier. In: Multimedia Tools and Applications. 2020 ; Vol. 79, No. 41-42. pp. 30653-30667.

Bibtex

@article{4e3ff3a0410c42a3ac08162c05cd8802,
title = "Human action recognition using deep rule-based classifier",
abstract = "In recent years, numerous techniques have been proposed for human activity recognition (HAR) from images and videos. These techniques can be divided into two major categories: handcrafted and deep learning. Deep Learning-based models have produced remarkable results for HAR. However, these models have several shortcomings, such as the requirement for a massive amount of training data, lack of transparency, offline nature, and poor interpretability of their internal parameters. In this paper, a new approach for HAR is proposed, which consists of an interpretable, self-evolving, and self-organizing set of 0-order If...THEN rules. This approach is entirely data-driven, and non-parametric; thus, prototypes are identified automatically during the training process. To demonstrate the effectiveness of the proposed method, a set of high-level features is obtained using a pre-trained deep convolution neural network model, and a recently introduced deep rule-based classifier is applied for classification. Experiments are performed on a challenging benchmark dataset UCF50; results confirmed that the proposed approach outperforms state-of-the-art methods. In addition to this, an ablation study is conducted to demonstrate the efficacy of the proposed approach by comparing the performance of our DRB classifier with four state-of-the-art classifiers. This analysis revealed that the DRB classifier could perform better than state-of-the-art classifiers, even with limited training samples.",
author = "Allah Bux and Xiaowei Gu and Plamen Angelov and Zulfiqar Habib",
note = "The final publication is available at Springer via http://dx.doi.org/10.1007/s11042-020-09381-9",
year = "2020",
month = nov,
day = "1",
doi = "10.1007/s11042-020-09381-9",
language = "English",
volume = "79",
pages = "30653--30667",
journal = "Multimedia Tools and Applications",
issn = "1380-7501",
publisher = "Springer Netherlands",
number = "41-42",

}

RIS

TY - JOUR

T1 - Human action recognition using deep rule-based classifier

AU - Bux, Allah

AU - Gu, Xiaowei

AU - Angelov, Plamen

AU - Habib, Zulfiqar

N1 - The final publication is available at Springer via http://dx.doi.org/10.1007/s11042-020-09381-9

PY - 2020/11/1

Y1 - 2020/11/1

N2 - In recent years, numerous techniques have been proposed for human activity recognition (HAR) from images and videos. These techniques can be divided into two major categories: handcrafted and deep learning. Deep Learning-based models have produced remarkable results for HAR. However, these models have several shortcomings, such as the requirement for a massive amount of training data, lack of transparency, offline nature, and poor interpretability of their internal parameters. In this paper, a new approach for HAR is proposed, which consists of an interpretable, self-evolving, and self-organizing set of 0-order If...THEN rules. This approach is entirely data-driven, and non-parametric; thus, prototypes are identified automatically during the training process. To demonstrate the effectiveness of the proposed method, a set of high-level features is obtained using a pre-trained deep convolution neural network model, and a recently introduced deep rule-based classifier is applied for classification. Experiments are performed on a challenging benchmark dataset UCF50; results confirmed that the proposed approach outperforms state-of-the-art methods. In addition to this, an ablation study is conducted to demonstrate the efficacy of the proposed approach by comparing the performance of our DRB classifier with four state-of-the-art classifiers. This analysis revealed that the DRB classifier could perform better than state-of-the-art classifiers, even with limited training samples.

AB - In recent years, numerous techniques have been proposed for human activity recognition (HAR) from images and videos. These techniques can be divided into two major categories: handcrafted and deep learning. Deep Learning-based models have produced remarkable results for HAR. However, these models have several shortcomings, such as the requirement for a massive amount of training data, lack of transparency, offline nature, and poor interpretability of their internal parameters. In this paper, a new approach for HAR is proposed, which consists of an interpretable, self-evolving, and self-organizing set of 0-order If...THEN rules. This approach is entirely data-driven, and non-parametric; thus, prototypes are identified automatically during the training process. To demonstrate the effectiveness of the proposed method, a set of high-level features is obtained using a pre-trained deep convolution neural network model, and a recently introduced deep rule-based classifier is applied for classification. Experiments are performed on a challenging benchmark dataset UCF50; results confirmed that the proposed approach outperforms state-of-the-art methods. In addition to this, an ablation study is conducted to demonstrate the efficacy of the proposed approach by comparing the performance of our DRB classifier with four state-of-the-art classifiers. This analysis revealed that the DRB classifier could perform better than state-of-the-art classifiers, even with limited training samples.

U2 - 10.1007/s11042-020-09381-9

DO - 10.1007/s11042-020-09381-9

M3 - Journal article

VL - 79

SP - 30653

EP - 30667

JO - Multimedia Tools and Applications

JF - Multimedia Tools and Applications

SN - 1380-7501

IS - 41-42

ER -