Details
| Original language | English |
|---|---|
| Article number | 7519028 |
| Pages (from-to) | 1364-1368 |
| Number of pages | 5 |
| Journal | IEEE Geoscience and Remote Sensing Letters |
| Volume | 13 |
| Issue number | 9 |
| Early online date | 22 Jul 2016 |
| Publication status | Published - Sept 2016 |
Abstract
This letter evaluates metaheuristics for the supervised parameter tuning of multiresolution-region-growing segmentation. Three groups of metaheuristics are tested in terms of convergence speed and solution quality. Generalized pattern search, mesh adaptive direct search, and Nelder-Mead represent the single-solution group. Differential evolution (DE) represents the population group. DE followed by each of the aforementioned single-solution metaheuristics represents the hybrid metaheuristic group. This letter reveals that the optimization objective functions typically have countless local minima, many of them leading to very poor solutions. Experiments on three data sets demonstrated that single-solution-based methods often lead to a solution with unacceptable quality. DE was less susceptible to be stuck in local minima when compared to single-solution methods, but it was slower in reaching the minima. Moreover, hybrid methods presented the best tradeoff between accuracy and convergence speed.
Keywords
- Image segmentation, optimization methods, parameter tuning
ASJC Scopus subject areas
- Earth and Planetary Sciences(all)
- Geotechnical Engineering and Engineering Geology
- Engineering(all)
- Electrical and Electronic Engineering
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: IEEE Geoscience and Remote Sensing Letters, Vol. 13, No. 9, 7519028, 09.2016, p. 1364-1368.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - Metaheuristics for Supervised Parameter Tuning of Multiresolution Segmentation
AU - Ayma Quirita, Victor Andres
AU - Achanccaray Diaz, Pedro
AU - Feitosa, Raul Q.
AU - Happ, Patrick N.
AU - Costa, Gilson A.O.P.
AU - Klinger, Tobias
AU - Heipke, Christian
PY - 2016/9
Y1 - 2016/9
N2 - This letter evaluates metaheuristics for the supervised parameter tuning of multiresolution-region-growing segmentation. Three groups of metaheuristics are tested in terms of convergence speed and solution quality. Generalized pattern search, mesh adaptive direct search, and Nelder-Mead represent the single-solution group. Differential evolution (DE) represents the population group. DE followed by each of the aforementioned single-solution metaheuristics represents the hybrid metaheuristic group. This letter reveals that the optimization objective functions typically have countless local minima, many of them leading to very poor solutions. Experiments on three data sets demonstrated that single-solution-based methods often lead to a solution with unacceptable quality. DE was less susceptible to be stuck in local minima when compared to single-solution methods, but it was slower in reaching the minima. Moreover, hybrid methods presented the best tradeoff between accuracy and convergence speed.
AB - This letter evaluates metaheuristics for the supervised parameter tuning of multiresolution-region-growing segmentation. Three groups of metaheuristics are tested in terms of convergence speed and solution quality. Generalized pattern search, mesh adaptive direct search, and Nelder-Mead represent the single-solution group. Differential evolution (DE) represents the population group. DE followed by each of the aforementioned single-solution metaheuristics represents the hybrid metaheuristic group. This letter reveals that the optimization objective functions typically have countless local minima, many of them leading to very poor solutions. Experiments on three data sets demonstrated that single-solution-based methods often lead to a solution with unacceptable quality. DE was less susceptible to be stuck in local minima when compared to single-solution methods, but it was slower in reaching the minima. Moreover, hybrid methods presented the best tradeoff between accuracy and convergence speed.
KW - Image segmentation
KW - optimization methods
KW - parameter tuning
UR - http://www.scopus.com/inward/record.url?scp=84979247258&partnerID=8YFLogxK
U2 - 10.1109/LGRS.2016.2586499
DO - 10.1109/LGRS.2016.2586499
M3 - Article
AN - SCOPUS:84979247258
VL - 13
SP - 1364
EP - 1368
JO - IEEE Geoscience and Remote Sensing Letters
JF - IEEE Geoscience and Remote Sensing Letters
SN - 1545-598X
IS - 9
M1 - 7519028
ER -