This Matlab Toolbox provides a set of functions to perform segmentation on scanning tunnel microsocpy images. The methods and algorithms are described in the paper
-K. Bui, et al. "Segmentation of Scanning Tunnel Microscopy Images Using Variational Methods and Empirical Wavelets", 2019
To cite this paper,
@Article{Bui2019,
author="Bui, Kevin
and Fauman, Jacob
and Kes, David
and Torres Mandiola, Leticia
and Ciomaga, Adina
and Salazar, Ricardo
and Bertozzi, Andrea L.
and Gilles, J{\'e}r{\^o}me
and Goronzy, Dominic P.
and Guttentag, Andrew I.
and Weiss, Paul S.",
title="Segmentation of scanning tunneling microscopy images using variational methods and empirical wavelets",
journal="Pattern Analysis and Applications",
year="2019",
month="May",
day="05",
issn="1433-755X",
doi="10.1007/s10044-019-00824-0",
url="https://doi.org/10.1007/s10044-019-00824-0"
}
This toolbox is freely distributed and can be used without any charges for research purposes, but please cite our papers when publishing your works. The code should not be used for commercial purposes.
For any questions, please contact kbui1993@gmail.com.
In the development of this toolbox, we require the following toolboxes developed by people outside the research group:
-Elad's Pseudo-Polar FFT toolbox: http://cs.techion.ac.il/~elad/software/ -Gilles' EWT toolbox: https://www.mathworks.com/matlabcentral/fileexchange/42141-empirical-wavelet-transforms
For the convenience of the user, the outside toolboxes are included in the Utilities subfolder in the Texture Segmentation folder. We do not claim that we contribute to any programming of the outside toolboxes.
This toolbox is organized as follows:
Microscopy Codes | |-Cartoon Segmentation: functions for cartoon segmentation |-Cartoon+Texture: functions for cartoon+texture decomposition |-Images: contains the images used in the paper |-Segmentation Framework: contains a script to run segmentation framework on an image |-Test: contains scripts to run basic tests on the images in Images folder |-Texture Segmentation: functions for texture segmentation |-Utilities: contains outside toolboxes
The Test folders contain scripts to reproduce the segmentation results of the images in our paper. The entire_cartoon_script.m and the entire_texture_script.m produce the cartoon and texture segmentation results, respectively, for all sixteen images in the Image folder. There is a single version if the user wants to run the segmentation algorithm onto one image. The single version could be modified if the user wants to run it on his or her own image. On the other hand, we have segmentation.m in Segmentation Framework folder where the user can specify the image and the parameters in the script and run the whole framework algorithm.