Vectorized data combination and binary search oriented reweight for CPU-GPU based real-time 3D ball tracking

Ziwei Deng; Yilin Hou; Xina Cheng; Takeshi Ikenaga

doi:10.1007/978-3-319-77383-4_50

Vectorized data combination and binary search oriented reweight for CPU-GPU based real-time 3D ball tracking

Ziwei Deng^*, Yilin Hou, Xina Cheng, Takeshi Ikenaga

^*Corresponding author for this work

Graduate School of Information, Production and Systems

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

1 Citation (Scopus)

Abstract

3D ball tracking is of great significance to sports analysis, which can be utilized to applications such as TV contents and tactic analysis. Some applications require real-time implementation, but a highly accurate tracking algorithm is usually time-consuming. This paper proposes a CPU-GPU platform based particle filter for multi-view ball tracking, including 2 proposals: vectorized mask data combination and binary search oriented reweight. The vectorized masks data combination unites HSV mask and inter-frame subtraction mask into one to reduce memory access time. The binary search oriented reweight helps getting and saving reweighted data with low complexity which could directly be used for binary search. The proposed methods are evaluated by both tracking accuracy and execution time. Experiment is based on GPU, the AMD R9 Fury, and compared to the serial implementation on CPU. The tracking accuracy keeps the same, while the execution time is reduced by a factor of 13.

Original language	English
Title of host publication	Advances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers
Editors	Bing Zeng, Hongliang Li, Qingming Huang, Abdulmotaleb El Saddik, Shuqiang Jiang, Xiaopeng Fan
Publisher	Springer Verlag
Pages	508-516
Number of pages	9
ISBN (Print)	9783319773827
DOIs	https://doi.org/10.1007/978-3-319-77383-4_50
Publication status	Published - 2018
Event	18th Pacific-Rim Conference on Multimedia, PCM 2017 - Harbin, China Duration: 2017 Sept 28 → 2017 Sept 29

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10736 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	18th Pacific-Rim Conference on Multimedia, PCM 2017
Country/Territory	China
City	Harbin
Period	17/9/28 → 17/9/29

Keywords

3D ball tracking
GPU acceleration
Heterogeneous computing
OpenCL
Particle filter
Sports analysis

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-319-77383-4_50

Cite this

Deng, Z., Hou, Y., Cheng, X., & Ikenaga, T. (2018). Vectorized data combination and binary search oriented reweight for CPU-GPU based real-time 3D ball tracking. In B. Zeng, H. Li, Q. Huang, A. El Saddik, S. Jiang, & X. Fan (Eds.), Advances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers (pp. 508-516). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10736 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-77383-4_50

Vectorized data combination and binary search oriented reweight for CPU-GPU based real-time 3D ball tracking. / Deng, Ziwei; Hou, Yilin; Cheng, Xina et al.
Advances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers. ed. / Bing Zeng; Hongliang Li; Qingming Huang; Abdulmotaleb El Saddik; Shuqiang Jiang; Xiaopeng Fan. Springer Verlag, 2018. p. 508-516 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10736 LNCS).

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

Deng, Z, Hou, Y, Cheng, X & Ikenaga, T 2018, Vectorized data combination and binary search oriented reweight for CPU-GPU based real-time 3D ball tracking. in B Zeng, H Li, Q Huang, A El Saddik, S Jiang & X Fan (eds), Advances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10736 LNCS, Springer Verlag, pp. 508-516, 18th Pacific-Rim Conference on Multimedia, PCM 2017, Harbin, China, 17/9/28. https://doi.org/10.1007/978-3-319-77383-4_50

Deng Z, Hou Y, Cheng X, Ikenaga T. Vectorized data combination and binary search oriented reweight for CPU-GPU based real-time 3D ball tracking. In Zeng B, Li H, Huang Q, El Saddik A, Jiang S, Fan X, editors, Advances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers. Springer Verlag. 2018. p. 508-516. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-77383-4_50

Deng, Ziwei ; Hou, Yilin ; Cheng, Xina et al. / Vectorized data combination and binary search oriented reweight for CPU-GPU based real-time 3D ball tracking. Advances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers. editor / Bing Zeng ; Hongliang Li ; Qingming Huang ; Abdulmotaleb El Saddik ; Shuqiang Jiang ; Xiaopeng Fan. Springer Verlag, 2018. pp. 508-516 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{dc24017ecbae4a449668e4e1be7c18c1,

title = "Vectorized data combination and binary search oriented reweight for CPU-GPU based real-time 3D ball tracking",

abstract = "3D ball tracking is of great significance to sports analysis, which can be utilized to applications such as TV contents and tactic analysis. Some applications require real-time implementation, but a highly accurate tracking algorithm is usually time-consuming. This paper proposes a CPU-GPU platform based particle filter for multi-view ball tracking, including 2 proposals: vectorized mask data combination and binary search oriented reweight. The vectorized masks data combination unites HSV mask and inter-frame subtraction mask into one to reduce memory access time. The binary search oriented reweight helps getting and saving reweighted data with low complexity which could directly be used for binary search. The proposed methods are evaluated by both tracking accuracy and execution time. Experiment is based on GPU, the AMD R9 Fury, and compared to the serial implementation on CPU. The tracking accuracy keeps the same, while the execution time is reduced by a factor of 13.",

keywords = "3D ball tracking, GPU acceleration, Heterogeneous computing, OpenCL, Particle filter, Sports analysis",

author = "Ziwei Deng and Yilin Hou and Xina Cheng and Takeshi Ikenaga",

note = "Funding Information: Acknowledgments. This work was supported by KAKENHI (16K 13006) and Waseda University Grant for Special Research Projects (2017K-263).; 18th Pacific-Rim Conference on Multimedia, PCM 2017 ; Conference date: 28-09-2017 Through 29-09-2017",

year = "2018",

doi = "10.1007/978-3-319-77383-4_50",

language = "English",

isbn = "9783319773827",

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

publisher = "Springer Verlag",

pages = "508--516",

editor = "Bing Zeng and Hongliang Li and Qingming Huang and {El Saddik}, Abdulmotaleb and Shuqiang Jiang and Xiaopeng Fan",

booktitle = "Advances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers",

}

TY - GEN

T1 - Vectorized data combination and binary search oriented reweight for CPU-GPU based real-time 3D ball tracking

AU - Deng, Ziwei

AU - Hou, Yilin

AU - Cheng, Xina

AU - Ikenaga, Takeshi

N1 - Funding Information: Acknowledgments. This work was supported by KAKENHI (16K 13006) and Waseda University Grant for Special Research Projects (2017K-263).

PY - 2018

Y1 - 2018

N2 - 3D ball tracking is of great significance to sports analysis, which can be utilized to applications such as TV contents and tactic analysis. Some applications require real-time implementation, but a highly accurate tracking algorithm is usually time-consuming. This paper proposes a CPU-GPU platform based particle filter for multi-view ball tracking, including 2 proposals: vectorized mask data combination and binary search oriented reweight. The vectorized masks data combination unites HSV mask and inter-frame subtraction mask into one to reduce memory access time. The binary search oriented reweight helps getting and saving reweighted data with low complexity which could directly be used for binary search. The proposed methods are evaluated by both tracking accuracy and execution time. Experiment is based on GPU, the AMD R9 Fury, and compared to the serial implementation on CPU. The tracking accuracy keeps the same, while the execution time is reduced by a factor of 13.

AB - 3D ball tracking is of great significance to sports analysis, which can be utilized to applications such as TV contents and tactic analysis. Some applications require real-time implementation, but a highly accurate tracking algorithm is usually time-consuming. This paper proposes a CPU-GPU platform based particle filter for multi-view ball tracking, including 2 proposals: vectorized mask data combination and binary search oriented reweight. The vectorized masks data combination unites HSV mask and inter-frame subtraction mask into one to reduce memory access time. The binary search oriented reweight helps getting and saving reweighted data with low complexity which could directly be used for binary search. The proposed methods are evaluated by both tracking accuracy and execution time. Experiment is based on GPU, the AMD R9 Fury, and compared to the serial implementation on CPU. The tracking accuracy keeps the same, while the execution time is reduced by a factor of 13.

KW - 3D ball tracking

KW - GPU acceleration

KW - Heterogeneous computing

KW - OpenCL

KW - Particle filter

KW - Sports analysis

UR - http://www.scopus.com/inward/record.url?scp=85047500066&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047500066&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-77383-4_50

DO - 10.1007/978-3-319-77383-4_50

M3 - Conference contribution

AN - SCOPUS:85047500066

SN - 9783319773827

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

SP - 508

EP - 516

BT - Advances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers

A2 - Zeng, Bing

A2 - Li, Hongliang

A2 - Huang, Qingming

A2 - El Saddik, Abdulmotaleb

A2 - Jiang, Shuqiang

A2 - Fan, Xiaopeng

PB - Springer Verlag

T2 - 18th Pacific-Rim Conference on Multimedia, PCM 2017

Y2 - 28 September 2017 through 29 September 2017

ER -

Vectorized data combination and binary search oriented reweight for CPU-GPU based real-time 3D ball tracking

Abstract

Publication series

Other

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this