Statistical tuning of Depth Map Algorithms
Egileak: Alejandro Hoyos and John Edgar Congote and Iñigo Barandiaran and Diego Acosta and O. Ruiz
Data: 22.08.2011
Abstract
In depth map generation, the settings of the algorithm parameters to yield an accurate disparity estimation are usually chosen empirically or based on unplanned experiments. A structured statistical approach including classical and exploratory data analyses on over 14000 images to measure the relative in uence of the parameters allows their tuning based on the number of bad pixels. The implemented methodology improves the performance of dense depth map algorithms. As a result of the statistical based tuning, the algorithm improves from 16.78% to 14.48% bad pixels in the Middlebury Stereo Evaluation Ranking Table. The performance is measured based on the distance of the algorithm results vs. the Ground Truth by Middlebury. Future work aims to achieve the tuning by using signicantly smaller data sets on fractional factorial and response surface design of experiments.
BIB_text
@Article {
author = {Alejandro Hoyos and John Edgar Congote and Iñigo Barandiaran and Diego Acosta and O. Ruiz},
title = {Statistical tuning of Depth Map Algorithms},
pages = {563-572},
volume = {6855},
keywds = {
Computer Science
}
abstract = {
In depth map generation, the settings of the algorithm parameters to yield an accurate disparity estimation are usually chosen empirically or based on unplanned experiments. A structured statistical approach including classical and exploratory data analyses on over 14000 images to measure the relative in uence of the parameters allows their tuning based on the number of bad pixels. The implemented methodology improves the performance of dense depth map algorithms. As a result of the statistical based tuning, the algorithm improves from 16.78% to 14.48% bad pixels in the Middlebury Stereo Evaluation Ranking Table. The performance is measured based on the distance of the algorithm results vs. the Ground Truth by Middlebury. Future work aims to achieve the tuning by using signicantly smaller data sets on fractional factorial and response surface design of experiments.
}
isbn = {978-3-642-23677-8},
date = {2011-08-22},
year = {2011},
}
author = {Alejandro Hoyos and John Edgar Congote and Iñigo Barandiaran and Diego Acosta and O. Ruiz},
title = {Statistical tuning of Depth Map Algorithms},
pages = {563-572},
volume = {6855},
keywds = {
Computer Science
}
abstract = {
In depth map generation, the settings of the algorithm parameters to yield an accurate disparity estimation are usually chosen empirically or based on unplanned experiments. A structured statistical approach including classical and exploratory data analyses on over 14000 images to measure the relative in uence of the parameters allows their tuning based on the number of bad pixels. The implemented methodology improves the performance of dense depth map algorithms. As a result of the statistical based tuning, the algorithm improves from 16.78% to 14.48% bad pixels in the Middlebury Stereo Evaluation Ranking Table. The performance is measured based on the distance of the algorithm results vs. the Ground Truth by Middlebury. Future work aims to achieve the tuning by using signicantly smaller data sets on fractional factorial and response surface design of experiments.
}
isbn = {978-3-642-23677-8},
date = {2011-08-22},
year = {2011},
}
