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

Research output: Chapter in Book/Report/Conference proceedingConference 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 languageEnglish
Title of host publicationAdvances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers
PublisherSpringer-Verlag
Pages508-516
Number of pages9
ISBN (Print)9783319773827
DOIs
Publication statusPublished - 2018 Jan 1
Event18th Pacific-Rim Conference on Multimedia, PCM 2017 - Harbin, China
Duration: 2017 Sep 282017 Sep 29

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10736 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other18th Pacific-Rim Conference on Multimedia, PCM 2017
CountryChina
CityHarbin
Period17/9/2817/9/29

Fingerprint

Binary search
Program processors
Masks
Ball
Mask
Real-time
Execution Time
Sports
Particle Filter
Subtraction
Low Complexity
Data storage equipment
Graphics processing unit
Experiments
Experiment

Keywords

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

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

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 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; Ikenaga, Takeshi.

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); Vol. 10736 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference 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 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 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)). https://doi.org/10.1007/978-3-319-77383-4_50
Deng, Ziwei ; Hou, Yilin ; Cheng, Xina ; Ikenaga, Takeshi. / 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. 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",
year = "2018",
month = "1",
day = "1",
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",
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

PY - 2018/1/1

Y1 - 2018/1/1

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

PB - Springer-Verlag

ER -