Skip to content
Back to outputs

Deep selective feature learning for action recognition

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

Standard

Deep selective feature learning for action recognition. / Li, Ziqiang; Ge, Yongxin; Feng, Jinyuan; Qin, Xiaolei; Yu, Jiaruo; Yu, Hui.

2020 IEEE International Conference on Multimedia and Expo, ICME 2020. Institute of Electrical and Electronics Engineers, 2020. 9102727 (Proceedings - IEEE International Conference on Multimedia and Expo; Vol. 2020-July).

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

Harvard

Li, Z, Ge, Y, Feng, J, Qin, X, Yu, J & Yu, H 2020, Deep selective feature learning for action recognition. in 2020 IEEE International Conference on Multimedia and Expo, ICME 2020., 9102727, Proceedings - IEEE International Conference on Multimedia and Expo, vol. 2020-July, Institute of Electrical and Electronics Engineers, 2020 IEEE International Conference on Multimedia and Expo, ICME 2020, London, United Kingdom, 6/07/20. https://doi.org/10.1109/ICME46284.2020.9102727

APA

Li, Z., Ge, Y., Feng, J., Qin, X., Yu, J., & Yu, H. (2020). Deep selective feature learning for action recognition. In 2020 IEEE International Conference on Multimedia and Expo, ICME 2020 [9102727] (Proceedings - IEEE International Conference on Multimedia and Expo; Vol. 2020-July). Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ICME46284.2020.9102727

Vancouver

Li Z, Ge Y, Feng J, Qin X, Yu J, Yu H. Deep selective feature learning for action recognition. In 2020 IEEE International Conference on Multimedia and Expo, ICME 2020. Institute of Electrical and Electronics Engineers. 2020. 9102727. (Proceedings - IEEE International Conference on Multimedia and Expo). https://doi.org/10.1109/ICME46284.2020.9102727

Author

Li, Ziqiang ; Ge, Yongxin ; Feng, Jinyuan ; Qin, Xiaolei ; Yu, Jiaruo ; Yu, Hui. / Deep selective feature learning for action recognition. 2020 IEEE International Conference on Multimedia and Expo, ICME 2020. Institute of Electrical and Electronics Engineers, 2020. (Proceedings - IEEE International Conference on Multimedia and Expo).

Bibtex

@inproceedings{1e19ec60bbf241b2bd7c33b4f9909dcb,
title = "Deep selective feature learning for action recognition",
abstract = "Soft-attention mechanism has attracted a lot of attention in recent years due to its ability to capture the most discriminative image features for understanding actions. However, soft-attention tends to focus on fine-grained parts on images and ignores global information, which can lead to totally wrong classification results. To address this issue, we propose a novel deep selective feature learning network (DSFNet), which can automatically learn the feature maps with both fine-grained and global information. Specially, DSFNet is designed to have the ability to learn to adjust the actions for feature map selection by maximizing the cumulative discounted rewards. Moreover, the DSFNet is an easy-to-use extension of state-of-the-art base architectures of multiple tasks. Extensive experiments show that the proposed method has achieved superior performance on two standard action recognition benchmarks across still images (PPMI) and videos (HMDB51).",
keywords = "Action recognition, Feature selection, Reinforcement learning",
author = "Ziqiang Li and Yongxin Ge and Jinyuan Feng and Xiaolei Qin and Jiaruo Yu and Hui Yu",
year = "2020",
month = jul,
day = "6",
doi = "10.1109/ICME46284.2020.9102727",
language = "English",
isbn = "978-1-7281-1332-6",
series = "Proceedings - IEEE International Conference on Multimedia and Expo",
publisher = "Institute of Electrical and Electronics Engineers",
booktitle = "2020 IEEE International Conference on Multimedia and Expo, ICME 2020",
note = "2020 IEEE International Conference on Multimedia and Expo, ICME 2020 ; Conference date: 06-07-2020 Through 10-07-2020",

}

RIS

TY - GEN

T1 - Deep selective feature learning for action recognition

AU - Li, Ziqiang

AU - Ge, Yongxin

AU - Feng, Jinyuan

AU - Qin, Xiaolei

AU - Yu, Jiaruo

AU - Yu, Hui

PY - 2020/7/6

Y1 - 2020/7/6

N2 - Soft-attention mechanism has attracted a lot of attention in recent years due to its ability to capture the most discriminative image features for understanding actions. However, soft-attention tends to focus on fine-grained parts on images and ignores global information, which can lead to totally wrong classification results. To address this issue, we propose a novel deep selective feature learning network (DSFNet), which can automatically learn the feature maps with both fine-grained and global information. Specially, DSFNet is designed to have the ability to learn to adjust the actions for feature map selection by maximizing the cumulative discounted rewards. Moreover, the DSFNet is an easy-to-use extension of state-of-the-art base architectures of multiple tasks. Extensive experiments show that the proposed method has achieved superior performance on two standard action recognition benchmarks across still images (PPMI) and videos (HMDB51).

AB - Soft-attention mechanism has attracted a lot of attention in recent years due to its ability to capture the most discriminative image features for understanding actions. However, soft-attention tends to focus on fine-grained parts on images and ignores global information, which can lead to totally wrong classification results. To address this issue, we propose a novel deep selective feature learning network (DSFNet), which can automatically learn the feature maps with both fine-grained and global information. Specially, DSFNet is designed to have the ability to learn to adjust the actions for feature map selection by maximizing the cumulative discounted rewards. Moreover, the DSFNet is an easy-to-use extension of state-of-the-art base architectures of multiple tasks. Extensive experiments show that the proposed method has achieved superior performance on two standard action recognition benchmarks across still images (PPMI) and videos (HMDB51).

KW - Action recognition

KW - Feature selection

KW - Reinforcement learning

UR - http://www.scopus.com/inward/record.url?scp=85090386398&partnerID=8YFLogxK

U2 - 10.1109/ICME46284.2020.9102727

DO - 10.1109/ICME46284.2020.9102727

M3 - Conference contribution

AN - SCOPUS:85090386398

SN - 978-1-7281-1332-6

T3 - Proceedings - IEEE International Conference on Multimedia and Expo

BT - 2020 IEEE International Conference on Multimedia and Expo, ICME 2020

PB - Institute of Electrical and Electronics Engineers

T2 - 2020 IEEE International Conference on Multimedia and Expo, ICME 2020

Y2 - 6 July 2020 through 10 July 2020

ER -

ID: 22579238