An efficient dense descriptor applied to 3D vision implemented on parallel computing

Dario I. Rosas-Miranda; Volodymyr I. Ponomaryov; Cesar M.A. Robles-Gonzalez

doi:10.1117/12.2303667

An efficient dense descriptor applied to 3D vision implemented on parallel computing

Dario I. Rosas-Miranda, Volodymyr I. Ponomaryov, Cesar M.A. Robles-Gonzalez

Escuela Superior de Ingeniería Mecánica y Eléctrica (ESIME), Unidad Culhuacán

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

Abstract

In this study, we introduce a novel local image descriptor, which is very efficient to compute densely. We also present an algorithm to compute dense depth maps from image pairs using designed descriptor. Novel descriptor is based on visual primitives and relations between them, namely coplanarity, cocolority, distance, and angle. Designed feature descriptor covers both geometric and appearance information. The depth map estimation performance is evaluated using the established bad matched pixel metric. An analysis of the feature descriptor employing a parallel programming paradigm is included to develop a possible real-time mode. This is performed with help of hardware based on multi-core processors and GPU platform, using a NVIDIA ® GeForce ® GT640 graphic card and Matlab over a PC with Windows 10.

Original language	English
Title of host publication	Real-Time Image and Video Processing 2018
Editors	Nasser Kehtarnavaz, Matthias F. Carlsohn
Publisher	SPIE
ISBN (Electronic)	9781510618510
DOIs	https://doi.org/10.1117/12.2303667
State	Published - 2018
Event	Real-Time Image and Video Processing 2018 - Orlando, United States Duration: 16 Apr 2018 → 17 Apr 2018

Publication series

Name	Proceedings of SPIE - The International Society for Optical Engineering
Volume	10670
ISSN (Print)	0277-786X
ISSN (Electronic)	1996-756X

Conference

Conference	Real-Time Image and Video Processing 2018
Country/Territory	United States
City	Orlando
Period	16/04/18 → 17/04/18

Keywords

Disparity Map
Features Descriptors
Image Descriptor
Parallel Computing

Access to Document

10.1117/12.2303667

Cite this

Rosas-Miranda, D. I., Ponomaryov, V. I., & Robles-Gonzalez, C. M. A. (2018). An efficient dense descriptor applied to 3D vision implemented on parallel computing. In N. Kehtarnavaz, & M. F. Carlsohn (Eds.), Real-Time Image and Video Processing 2018 Article 1067007 (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 10670). SPIE. https://doi.org/10.1117/12.2303667

@inproceedings{4c1a0e6eec38459992541cdff289e3be,

title = "An efficient dense descriptor applied to 3D vision implemented on parallel computing",

abstract = "In this study, we introduce a novel local image descriptor, which is very efficient to compute densely. We also present an algorithm to compute dense depth maps from image pairs using designed descriptor. Novel descriptor is based on visual primitives and relations between them, namely coplanarity, cocolority, distance, and angle. Designed feature descriptor covers both geometric and appearance information. The depth map estimation performance is evaluated using the established bad matched pixel metric. An analysis of the feature descriptor employing a parallel programming paradigm is included to develop a possible real-time mode. This is performed with help of hardware based on multi-core processors and GPU platform, using a NVIDIA {\textregistered} GeForce {\textregistered} GT640 graphic card and Matlab over a PC with Windows 10.",

keywords = "Disparity Map, Features Descriptors, Image Descriptor, Parallel Computing",

author = "Rosas-Miranda, {Dario I.} and Ponomaryov, {Volodymyr I.} and Robles-Gonzalez, {Cesar M.A.}",

note = "Publisher Copyright: {\textcopyright} 2018 SPIE. Downloading of the abstract is permitted for personal use only.; Real-Time Image and Video Processing 2018 ; Conference date: 16-04-2018 Through 17-04-2018",

year = "2018",

doi = "10.1117/12.2303667",

language = "Ingl{\'e}s",

series = "Proceedings of SPIE - The International Society for Optical Engineering",

publisher = "SPIE",

editor = "Nasser Kehtarnavaz and Carlsohn, {Matthias F.}",

booktitle = "Real-Time Image and Video Processing 2018",

address = "Estados Unidos",

}

Rosas-Miranda, DI, Ponomaryov, VI & Robles-Gonzalez, CMA 2018, An efficient dense descriptor applied to 3D vision implemented on parallel computing. in N Kehtarnavaz & MF Carlsohn (eds), Real-Time Image and Video Processing 2018., 1067007, Proceedings of SPIE - The International Society for Optical Engineering, vol. 10670, SPIE, Real-Time Image and Video Processing 2018, Orlando, United States, 16/04/18. https://doi.org/10.1117/12.2303667

An efficient dense descriptor applied to 3D vision implemented on parallel computing. / Rosas-Miranda, Dario I.; Ponomaryov, Volodymyr I.; Robles-Gonzalez, Cesar M.A.
Real-Time Image and Video Processing 2018. ed. / Nasser Kehtarnavaz; Matthias F. Carlsohn. SPIE, 2018. 1067007 (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 10670).

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

TY - GEN

T1 - An efficient dense descriptor applied to 3D vision implemented on parallel computing

AU - Rosas-Miranda, Dario I.

AU - Ponomaryov, Volodymyr I.

AU - Robles-Gonzalez, Cesar M.A.

PY - 2018

Y1 - 2018

N2 - In this study, we introduce a novel local image descriptor, which is very efficient to compute densely. We also present an algorithm to compute dense depth maps from image pairs using designed descriptor. Novel descriptor is based on visual primitives and relations between them, namely coplanarity, cocolority, distance, and angle. Designed feature descriptor covers both geometric and appearance information. The depth map estimation performance is evaluated using the established bad matched pixel metric. An analysis of the feature descriptor employing a parallel programming paradigm is included to develop a possible real-time mode. This is performed with help of hardware based on multi-core processors and GPU platform, using a NVIDIA ® GeForce ® GT640 graphic card and Matlab over a PC with Windows 10.

AB - In this study, we introduce a novel local image descriptor, which is very efficient to compute densely. We also present an algorithm to compute dense depth maps from image pairs using designed descriptor. Novel descriptor is based on visual primitives and relations between them, namely coplanarity, cocolority, distance, and angle. Designed feature descriptor covers both geometric and appearance information. The depth map estimation performance is evaluated using the established bad matched pixel metric. An analysis of the feature descriptor employing a parallel programming paradigm is included to develop a possible real-time mode. This is performed with help of hardware based on multi-core processors and GPU platform, using a NVIDIA ® GeForce ® GT640 graphic card and Matlab over a PC with Windows 10.

KW - Disparity Map

KW - Features Descriptors

KW - Image Descriptor

KW - Parallel Computing

UR - http://www.scopus.com/inward/record.url?scp=85049585330&partnerID=8YFLogxK

U2 - 10.1117/12.2303667

DO - 10.1117/12.2303667

M3 - Contribución a la conferencia

AN - SCOPUS:85049585330

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Real-Time Image and Video Processing 2018

A2 - Kehtarnavaz, Nasser

A2 - Carlsohn, Matthias F.

PB - SPIE

T2 - Real-Time Image and Video Processing 2018

Y2 - 16 April 2018 through 17 April 2018

ER -

An efficient dense descriptor applied to 3D vision implemented on parallel computing

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this