A versatile stereo implementation on commodity graphics hardware

Ruigang Yang; Marc Pollefeys

doi:10.1016/j.rti.2005.04.002

A versatile stereo implementation on commodity graphics hardware

Ruigang Yang, Marc Pollefeys

Computer Science

Research output: Contribution to journal › Article › peer-review

48 Scopus citations

Abstract

This paper presents a detailed description of a real-time correlation-based stereo algorithm running completely on the graphics processing unit (GPU). This is important since it allows to free up the main processor for other tasks including high-level interpretation of the stereo results. We first introduce a two-view stereo algorithm that includes some advanced features such as adaptive windows and cross-checking. Then we extend it using a plane-sweep approach to allow multiple frames without rectification. By taking advantage of advanced features of recent GPUs the proposed algorithm runs in real-time. Our implementation running on an ATI Radeon 9800 graphics card achieves up to 289 million disparity evaluations per second including all the overhead to download images and read-back the disparity map, which is several times faster than commercially available CPU-based implementations.

Original language	English
Pages (from-to)	7-18
Number of pages	12
Journal	Real-Time Imaging
Volume	11
Issue number	1
DOIs	https://doi.org/10.1016/j.rti.2005.04.002
State	Published - Feb 2005

Bibliographical note

Funding Information:
The authors would like to thank ATI and NVIDIA for their generous hardware donations and technical support. The work is supported in part by NSF grant IIS-0313047, IIS-0448185, and the office of research at the University of Kentucky.

ASJC Scopus subject areas

Signal Processing
Computer Vision and Pattern Recognition
Electrical and Electronic Engineering

Access to Document

10.1016/j.rti.2005.04.002

Cite this

@article{8d83f3367d754e25b371ff01e6cd7510,

title = "A versatile stereo implementation on commodity graphics hardware",

abstract = "This paper presents a detailed description of a real-time correlation-based stereo algorithm running completely on the graphics processing unit (GPU). This is important since it allows to free up the main processor for other tasks including high-level interpretation of the stereo results. We first introduce a two-view stereo algorithm that includes some advanced features such as adaptive windows and cross-checking. Then we extend it using a plane-sweep approach to allow multiple frames without rectification. By taking advantage of advanced features of recent GPUs the proposed algorithm runs in real-time. Our implementation running on an ATI Radeon 9800 graphics card achieves up to 289 million disparity evaluations per second including all the overhead to download images and read-back the disparity map, which is several times faster than commercially available CPU-based implementations.",

author = "Ruigang Yang and Marc Pollefeys",

note = "Funding Information: The authors would like to thank ATI and NVIDIA for their generous hardware donations and technical support. The work is supported in part by NSF grant IIS-0313047, IIS-0448185, and the office of research at the University of Kentucky.",

year = "2005",

month = feb,

doi = "10.1016/j.rti.2005.04.002",

language = "English",

volume = "11",

pages = "7--18",

number = "1",

}

TY - JOUR

T1 - A versatile stereo implementation on commodity graphics hardware

AU - Yang, Ruigang

AU - Pollefeys, Marc

N1 - Funding Information: The authors would like to thank ATI and NVIDIA for their generous hardware donations and technical support. The work is supported in part by NSF grant IIS-0313047, IIS-0448185, and the office of research at the University of Kentucky.

PY - 2005/2

Y1 - 2005/2

N2 - This paper presents a detailed description of a real-time correlation-based stereo algorithm running completely on the graphics processing unit (GPU). This is important since it allows to free up the main processor for other tasks including high-level interpretation of the stereo results. We first introduce a two-view stereo algorithm that includes some advanced features such as adaptive windows and cross-checking. Then we extend it using a plane-sweep approach to allow multiple frames without rectification. By taking advantage of advanced features of recent GPUs the proposed algorithm runs in real-time. Our implementation running on an ATI Radeon 9800 graphics card achieves up to 289 million disparity evaluations per second including all the overhead to download images and read-back the disparity map, which is several times faster than commercially available CPU-based implementations.

AB - This paper presents a detailed description of a real-time correlation-based stereo algorithm running completely on the graphics processing unit (GPU). This is important since it allows to free up the main processor for other tasks including high-level interpretation of the stereo results. We first introduce a two-view stereo algorithm that includes some advanced features such as adaptive windows and cross-checking. Then we extend it using a plane-sweep approach to allow multiple frames without rectification. By taking advantage of advanced features of recent GPUs the proposed algorithm runs in real-time. Our implementation running on an ATI Radeon 9800 graphics card achieves up to 289 million disparity evaluations per second including all the overhead to download images and read-back the disparity map, which is several times faster than commercially available CPU-based implementations.

UR - http://www.scopus.com/inward/record.url?scp=19744378665&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=19744378665&partnerID=8YFLogxK

U2 - 10.1016/j.rti.2005.04.002

DO - 10.1016/j.rti.2005.04.002

M3 - Article

AN - SCOPUS:19744378665

SN - 1077-2014

VL - 11

SP - 7

EP - 18

JO - Real-Time Imaging

JF - Real-Time Imaging

IS - 1

ER -

A versatile stereo implementation on commodity graphics hardware

Abstract

Bibliographical note

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this