GPU accelerated likelihoods for stereo-based articulated tracking

Rune Møllegaard Friborg; Søren Hauberg; Kenny Erleben

doi:10.1007/978-3-642-35740-4_28

GPU accelerated likelihoods for stereo-based articulated tracking

Rune Møllegaard Friborg, Søren Hauberg, Kenny Erleben

Department of Computer Science

1 Citation (Scopus)

Abstract

For many years articulated tracking has been an active research topic in the computer vision community. While working solutions have been suggested, computational time is still problematic. We present a GPU implementation of a ray-casting based likelihood model that is orders of magnitude faster than a traditional CPU implementation. We explain the non-intuitive steps required to attain an optimized GPU implementation, where the dominant part is to hide the memory latency effectively. Benchmarks show that computations which previously required
several minutes, are now performed in few seconds.

Original language	English
Title of host publication	Trends and Topics in Computer Vision : ECCV 2010 Workshops, Heraklion, Crete, Greece, September 10-11, 2010, Revised Selected Papers, Part II
Editors	Kiriakos N. Kutulakos
Number of pages	13
Volume	Part II
Publisher	Springer
Publication date	2012
Pages	359-371
ISBN (Print)	978-3-642-35739-8
ISBN (Electronic)	978-3-642-35740-4
DOIs	https://doi.org/10.1007/978-3-642-35740-4_28
Publication status	Published - 2012
Event	Workshop on Computer Vision on GPUs - Heraklion, Greece Duration: 10 Sept 2010 → 11 Sept 2010

Workshop

Workshop	Workshop on Computer Vision on GPUs
Country/Territory	Greece
City	Heraklion
Period	10/09/2010 → 11/09/2010

Series	Lecture notes in computer science
Volume	6554
ISSN	0302-9743

Access to Document

10.1007/978-3-642-35740-4_28

https://link.springer.com/content/pdf/10.1007/978-3-642-35740-4_28.pdf

Cite this

Friborg, R. M., Hauberg, S., & Erleben, K. (2012). GPU accelerated likelihoods for stereo-based articulated tracking. In K. N. Kutulakos (Ed.), Trends and Topics in Computer Vision: ECCV 2010 Workshops, Heraklion, Crete, Greece, September 10-11, 2010, Revised Selected Papers, Part II (Vol. Part II, pp. 359-371). Springer. Lecture notes in computer science Vol. 6554 https://doi.org/10.1007/978-3-642-35740-4_28

GPU accelerated likelihoods for stereo-based articulated tracking. / Friborg, Rune Møllegaard; Hauberg, Søren; Erleben, Kenny.

Trends and Topics in Computer Vision: ECCV 2010 Workshops, Heraklion, Crete, Greece, September 10-11, 2010, Revised Selected Papers, Part II. ed. / Kiriakos N. Kutulakos. Vol. Part II Springer, 2012. p. 359-371 (Lecture notes in computer science, Vol. 6554).

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

Friborg, RM, Hauberg, S & Erleben, K 2012, GPU accelerated likelihoods for stereo-based articulated tracking. in KN Kutulakos (ed.), Trends and Topics in Computer Vision: ECCV 2010 Workshops, Heraklion, Crete, Greece, September 10-11, 2010, Revised Selected Papers, Part II. vol. Part II, Springer, Lecture notes in computer science, vol. 6554, pp. 359-371, Workshop on Computer Vision on GPUs, Heraklion, Greece, 10/09/2010. https://doi.org/10.1007/978-3-642-35740-4_28

Friborg RM, Hauberg S, Erleben K. GPU accelerated likelihoods for stereo-based articulated tracking. In Kutulakos KN, editor, Trends and Topics in Computer Vision: ECCV 2010 Workshops, Heraklion, Crete, Greece, September 10-11, 2010, Revised Selected Papers, Part II. Vol. Part II. Springer. 2012. p. 359-371. (Lecture notes in computer science, Vol. 6554). doi: 10.1007/978-3-642-35740-4_28

Friborg, Rune Møllegaard ; Hauberg, Søren ; Erleben, Kenny. / GPU accelerated likelihoods for stereo-based articulated tracking. Trends and Topics in Computer Vision: ECCV 2010 Workshops, Heraklion, Crete, Greece, September 10-11, 2010, Revised Selected Papers, Part II. editor / Kiriakos N. Kutulakos. Vol. Part II Springer, 2012. pp. 359-371 (Lecture notes in computer science, Vol. 6554).

@inproceedings{53a8a0d0c1d111df825b000ea68e967b,

title = "GPU accelerated likelihoods for stereo-based articulated tracking",

abstract = "For many years articulated tracking has been an active research topic in the computer vision community. While working solutions have been suggested, computational time is still problematic. We present a GPU implementation of a ray-casting based likelihood model that is orders of magnitude faster than a traditional CPU implementation. We explain the non-intuitive steps required to attain an optimized GPU implementation, where the dominant part is to hide the memory latency effectively. Benchmarks show that computations which previously requiredseveral minutes, are now performed in few seconds.",

author = "Friborg, {Rune M{\o}llegaard} and S{\o}ren Hauberg and Kenny Erleben",

year = "2012",

doi = "10.1007/978-3-642-35740-4_28",

language = "English",

isbn = "978-3-642-35739-8",

volume = "Part II",

series = "Lecture notes in computer science",

publisher = "Springer",

pages = "359--371",

editor = "Kutulakos, {Kiriakos N.}",

booktitle = "Trends and Topics in Computer Vision",

note = "Workshop on Computer Vision on GPUs ; Conference date: 10-09-2010 Through 11-09-2010",

}

TY - GEN

T1 - GPU accelerated likelihoods for stereo-based articulated tracking

AU - Friborg, Rune Møllegaard

AU - Hauberg, Søren

AU - Erleben, Kenny

PY - 2012

Y1 - 2012

N2 - For many years articulated tracking has been an active research topic in the computer vision community. While working solutions have been suggested, computational time is still problematic. We present a GPU implementation of a ray-casting based likelihood model that is orders of magnitude faster than a traditional CPU implementation. We explain the non-intuitive steps required to attain an optimized GPU implementation, where the dominant part is to hide the memory latency effectively. Benchmarks show that computations which previously requiredseveral minutes, are now performed in few seconds.

AB - For many years articulated tracking has been an active research topic in the computer vision community. While working solutions have been suggested, computational time is still problematic. We present a GPU implementation of a ray-casting based likelihood model that is orders of magnitude faster than a traditional CPU implementation. We explain the non-intuitive steps required to attain an optimized GPU implementation, where the dominant part is to hide the memory latency effectively. Benchmarks show that computations which previously requiredseveral minutes, are now performed in few seconds.

U2 - 10.1007/978-3-642-35740-4_28

DO - 10.1007/978-3-642-35740-4_28

M3 - Article in proceedings

SN - 978-3-642-35739-8

VL - Part II

T3 - Lecture notes in computer science

SP - 359

EP - 371

BT - Trends and Topics in Computer Vision

A2 - Kutulakos, Kiriakos N.

PB - Springer

T2 - Workshop on Computer Vision on GPUs

Y2 - 10 September 2010 through 11 September 2010

ER -