This repository contains all the code and results of the paper entitled "Neuro-inspired edge feature fusion using Choquet integrals" published in the journal Information Sciences.
Paper link to Publisher site.
authored by
C. Marco-Detchart, G. Lucca, C. Lopez-Molina, L. De Miguel, G. P. Dimuro and H. Bustince.
Corresponding author: Cedric Marco-Detchart (cedric.marco@unavarra.es)
The article can be found in PDF format here, as the file "Neuro-inspired-edge-feature-fusion-using-Choquet-integrals.pdf".
Prerequisites: The KITT (Kermit Image Toolkit) collection is needed for this project to work. The folders needed are loaded in "setup.m" file, where paths can be configured. Please download the files from KITT repository.
To execute the experiment, run the file "superLauncher.m" where the OS (win, mac, linux) must be chosen in order to build the correct path syntax.
The parameters configuration is located in the file "infoMaker.m". The source and data paths must also be configured according to your folders location.
Each one of the phases of the experiment is located in one file, as follows:
-
"smMaker.m" contains the procedure to apply a smoothing technique to an image, taking all the parameters from configuration file.
-
"ftMaker.m" is responsible for extracting the feature images based on given parameters.
-
"bdryMaker.m" takes a feature image and extracts boundary image so that they can be compared to ground truth.
-
"cpMaker.m" computes the comparison of a boundary image with its ground truth giving statistical results (Prec, Rec and F_0.5 measure).
-
"cpCollecter.m" takes individual statistical results and collects them all in order to have a global result of the dataset
-
"aioMakerClassic.m" contains all of the instructions of the previous files in order to process the dataset images with the classical methods to whom we compare.
-
"aioMaker.m" contains all of the instructions of the previous files in one script.
-
"aioMakerSelection.m" permit to execute the experiment all at once with a selected list of specific parameters.
-
"README.md" this file.
You will find the file "Results-Choquet-generalizations-edges.xlsx" containing the complete experimental results done for testing our proposal. The file presents the Choquet generalization used, with its corresponding function/s along with the exponent of the power measure. For each combination a series of statistical results is shown (Prec, Rec and F_0.5 measure).
If you use this code and/or article in your research, please cite as:
@article{MARCODETCHART2021,
title = {Neuro-inspired edge feature fusion using Choquet integrals},
journal = {Information Sciences},
volume = {581},
pages = {740-754},
year = {2021},
issn = {0020-0255},
doi = {https://doi.org/10.1016/j.ins.2021.10.016},
url = {https://www.sciencedirect.com/science/article/pii/S0020025521010240},
author = {Cedric Marco-Detchart and Giancarlo Lucca and Carlos Lopez-Molina and Laura De Miguel and Graçaliz Pereira Dimuro and Humberto Bustince},
}