by Morales, S and Klette, R
Abstract:
Current stereo algorithms are capable to calculate accurate (as defined, e.g., by needs in vision-based driver assistance) dense disparity maps in real time. They have become the source of three-dimensional data for several indoor and outdoor applications. However, ground truth-based evaluation of such algorithms has been typically limited to data sets generated indoors in laboratories. In this paper we present a new approach to evaluate stereo algorithms using ground-truth over real world data sets. Ground truth is generated using range measurements acquired with a high-end laser range-finder. For evaluating as many points as possible in a given disparity map, we use two evaluation approaches: A direct comparison for those pixels with available range data, and a confidence measure for the remaining pixels. © 2011 Springer-Verlag Berlin Heidelberg.
Reference:
Ground truth evaluation of stereo algorithms for real world applications (Morales, S and Klette, R), In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), volume 6469 LNCS, 2011.
Bibtex Entry:
@inproceedings{morales2011groundapplications,
author = "Morales, S and Klette, R",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "152--162",
title = "Ground truth evaluation of stereo algorithms for real world applications",
volume = "6469 LNCS",
year = "2011",
abstract = "Current stereo algorithms are capable to calculate accurate (as defined, e.g., by needs in vision-based driver assistance) dense disparity maps in real time. They have become the source of three-dimensional data for several indoor and outdoor applications. However, ground truth-based evaluation of such algorithms has been typically limited to data sets generated indoors in laboratories. In this paper we present a new approach to evaluate stereo algorithms using ground-truth over real world data sets. Ground truth is generated using range measurements acquired with a high-end laser range-finder. For evaluating as many points as possible in a given disparity map, we use two evaluation approaches: A direct comparison for those pixels with available range data, and a confidence measure for the remaining pixels. © 2011 Springer-Verlag Berlin Heidelberg.",
doi = "10.1007/978-3-642-22819-3_16",
isbn = "9783642228186",
issn = "0302-9743",
eissn = "1611-3349",
issue = "PART 2",
keyword = "laser range finder",
keyword = "Performance evaluation",
keyword = "stereo algorithms",
language = "eng",
}