Assigning PLS based descriptors by SVM in action recognition

Jiayu Sheng; Biyun Sheng; Wankou Yang; Changyin Sun

doi:10.1007/978-3-319-23989-7_16

Assigning PLS based descriptors by SVM in action recognition

Jiayu Sheng
, Biyun Sheng
, Wankou Yang
, Changyin Sun

Southeast University, Nanjing

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

In this paper, we propose assigning PLS based descriptors by SVM to obtain the representations of human action videos. First, in addition to the spatially gradient orientation, we add spatio-temporal gradient statistic to generate the extended Histogram of Oriented Gradient (HOG). Second, different from requently-used cuboid descriptors in which Principal Component Analysis (PCA) is applied for dimension reduction, the proposed features utilize the Partial Least Squares (PLS) method for better performance. Then, we apply a multi-class SVM for assignment instead of assigning descriptors to the nearest (Euclidean distance) visual word in traditional Bag of Visual Words (BOVW) frame-work. Finally, the K-nearest neighbor algorithm is used to classify the histogram of visual words. The experimental results on the facial expression dataset and KTH human activity dataset validate the effectiveness of our proposed method.

Original language	English
Title of host publication	Intelligence Science and Big Data Engineering
Subtitle of host publication	Image and Video Data Engineering - 5th International Conference, IScIDE 2015, Revised Selected Papers
Editors	Xiaofei He, Zhi-Hua Zhou, Xinbo Gao, Zhi-Yong Liu, Yanning Zhang, Baochuan Fu, Fuyuan Hu, Zhancheng Zhang
Publisher	Springer Verlag
Pages	145-153
Number of pages	9
ISBN (Print)	9783319239873
DOIs	https://doi.org/10.1007/978-3-319-23989-7_16
State	Published - 2015
Externally published	Yes
Event	5th International Conference on Intelligence Science and Big Data Engineering, IScIDE 2015 - Suzhou, China Duration: 14 Jun 2015 → 16 Jun 2015

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9242
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	5th International Conference on Intelligence Science and Big Data Engineering, IScIDE 2015
Country/Territory	China
City	Suzhou
Period	14/06/15 → 16/06/15

Access to Document

10.1007/978-3-319-23989-7_16

Cite this

Sheng, J., Sheng, B., Yang, W., & Sun, C. (2015). Assigning PLS based descriptors by SVM in action recognition. In X. He, Z.-H. Zhou, X. Gao, Z.-Y. Liu, Y. Zhang, B. Fu, F. Hu, & Z. Zhang (Eds.), Intelligence Science and Big Data Engineering: Image and Video Data Engineering - 5th International Conference, IScIDE 2015, Revised Selected Papers (pp. 145-153). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9242). Springer Verlag. https://doi.org/10.1007/978-3-319-23989-7_16

Sheng, Jiayu ; Sheng, Biyun ; Yang, Wankou et al. / Assigning PLS based descriptors by SVM in action recognition. Intelligence Science and Big Data Engineering: Image and Video Data Engineering - 5th International Conference, IScIDE 2015, Revised Selected Papers. editor / Xiaofei He ; Zhi-Hua Zhou ; Xinbo Gao ; Zhi-Yong Liu ; Yanning Zhang ; Baochuan Fu ; Fuyuan Hu ; Zhancheng Zhang. Springer Verlag, 2015. pp. 145-153 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{9040778e256d4e06b487f15971cf2c05,

title = "Assigning PLS based descriptors by SVM in action recognition",

abstract = "In this paper, we propose assigning PLS based descriptors by SVM to obtain the representations of human action videos. First, in addition to the spatially gradient orientation, we add spatio-temporal gradient statistic to generate the extended Histogram of Oriented Gradient (HOG). Second, different from requently-used cuboid descriptors in which Principal Component Analysis (PCA) is applied for dimension reduction, the proposed features utilize the Partial Least Squares (PLS) method for better performance. Then, we apply a multi-class SVM for assignment instead of assigning descriptors to the nearest (Euclidean distance) visual word in traditional Bag of Visual Words (BOVW) frame-work. Finally, the K-nearest neighbor algorithm is used to classify the histogram of visual words. The experimental results on the facial expression dataset and KTH human activity dataset validate the effectiveness of our proposed method.",

author = "Jiayu Sheng and Biyun Sheng and Wankou Yang and Changyin Sun",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2015.; 5th International Conference on Intelligence Science and Big Data Engineering, IScIDE 2015 ; Conference date: 14-06-2015 Through 16-06-2015",

year = "2015",

doi = "10.1007/978-3-319-23989-7\_16",

language = "英语",

isbn = "9783319239873",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "145--153",

editor = "Xiaofei He and Zhi-Hua Zhou and Xinbo Gao and Zhi-Yong Liu and Yanning Zhang and Baochuan Fu and Fuyuan Hu and Zhancheng Zhang",

booktitle = "Intelligence Science and Big Data Engineering",

}

Sheng, J, Sheng, B, Yang, W & Sun, C 2015, Assigning PLS based descriptors by SVM in action recognition. in X He, Z-H Zhou, X Gao, Z-Y Liu, Y Zhang, B Fu, F Hu & Z Zhang (eds), Intelligence Science and Big Data Engineering: Image and Video Data Engineering - 5th International Conference, IScIDE 2015, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9242, Springer Verlag, pp. 145-153, 5th International Conference on Intelligence Science and Big Data Engineering, IScIDE 2015, Suzhou, China, 14/06/15. https://doi.org/10.1007/978-3-319-23989-7_16

Assigning PLS based descriptors by SVM in action recognition. / Sheng, Jiayu; Sheng, Biyun; Yang, Wankou et al.
Intelligence Science and Big Data Engineering: Image and Video Data Engineering - 5th International Conference, IScIDE 2015, Revised Selected Papers. ed. / Xiaofei He; Zhi-Hua Zhou; Xinbo Gao; Zhi-Yong Liu; Yanning Zhang; Baochuan Fu; Fuyuan Hu; Zhancheng Zhang. Springer Verlag, 2015. p. 145-153 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9242).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Assigning PLS based descriptors by SVM in action recognition

AU - Sheng, Jiayu

AU - Sheng, Biyun

AU - Yang, Wankou

AU - Sun, Changyin

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2015.

PY - 2015

Y1 - 2015

N2 - In this paper, we propose assigning PLS based descriptors by SVM to obtain the representations of human action videos. First, in addition to the spatially gradient orientation, we add spatio-temporal gradient statistic to generate the extended Histogram of Oriented Gradient (HOG). Second, different from requently-used cuboid descriptors in which Principal Component Analysis (PCA) is applied for dimension reduction, the proposed features utilize the Partial Least Squares (PLS) method for better performance. Then, we apply a multi-class SVM for assignment instead of assigning descriptors to the nearest (Euclidean distance) visual word in traditional Bag of Visual Words (BOVW) frame-work. Finally, the K-nearest neighbor algorithm is used to classify the histogram of visual words. The experimental results on the facial expression dataset and KTH human activity dataset validate the effectiveness of our proposed method.

AB - In this paper, we propose assigning PLS based descriptors by SVM to obtain the representations of human action videos. First, in addition to the spatially gradient orientation, we add spatio-temporal gradient statistic to generate the extended Histogram of Oriented Gradient (HOG). Second, different from requently-used cuboid descriptors in which Principal Component Analysis (PCA) is applied for dimension reduction, the proposed features utilize the Partial Least Squares (PLS) method for better performance. Then, we apply a multi-class SVM for assignment instead of assigning descriptors to the nearest (Euclidean distance) visual word in traditional Bag of Visual Words (BOVW) frame-work. Finally, the K-nearest neighbor algorithm is used to classify the histogram of visual words. The experimental results on the facial expression dataset and KTH human activity dataset validate the effectiveness of our proposed method.

UR - https://www.scopus.com/pages/publications/84951786437

U2 - 10.1007/978-3-319-23989-7_16

DO - 10.1007/978-3-319-23989-7_16

M3 - 会议稿件

AN - SCOPUS:84951786437

SN - 9783319239873

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 145

EP - 153

BT - Intelligence Science and Big Data Engineering

A2 - He, Xiaofei

A2 - Zhou, Zhi-Hua

A2 - Gao, Xinbo

A2 - Liu, Zhi-Yong

A2 - Zhang, Yanning

A2 - Fu, Baochuan

A2 - Hu, Fuyuan

A2 - Zhang, Zhancheng

PB - Springer Verlag

T2 - 5th International Conference on Intelligence Science and Big Data Engineering, IScIDE 2015

Y2 - 14 June 2015 through 16 June 2015

ER -

Sheng J, Sheng B, Yang W, Sun C. Assigning PLS based descriptors by SVM in action recognition. In He X, Zhou ZH, Gao X, Liu ZY, Zhang Y, Fu B, Hu F, Zhang Z, editors, Intelligence Science and Big Data Engineering: Image and Video Data Engineering - 5th International Conference, IScIDE 2015, Revised Selected Papers. Springer Verlag. 2015. p. 145-153. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-23989-7_16

Assigning PLS based descriptors by SVM in action recognition

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this