StepForest: Machine learning approach for segmenting glands in colon histology images using local intensity and texture features
A machine learning based Image Segmentation alrorithm, created for glands segmentation in colon histology images, can be modified for other image segmentation problems. This algorithm uses a novel hierarchical random forest apporach, where 3 levels of random forest beeen used for better segmentation.
For testing this algorithm, dataset of the GlaS@MICCAI'2015: Gland Segmentation Challenge Contest (https://warwick.ac.uk/fac/sci/dcs/research/tia/glascontest/) was used. Available to download under the 'Download' tab in the above mentioned website.
3rd Party Toolboxes/Codes Used (Governed by licenses provided by the respective authors):-
- haralickTextureFeatures by Rune Monzel (https://www.mathworks.com/matlabcentral/fileexchange/58769-haralicktexturefeatures)
- Stain Normalisation Toolbox for Matlab by Nicholas Trahearn and Adnan Khan of University of Warwick (https://warwick.ac.uk/fac/sci/dcs/research/tia/software/sntoolbox/) The source codes of these 3rd party toolboxes/codes are uploaded under the 'Toolboxes' Folder. Latest versions can be downloaded and license info can be obtained from the given websites
The research was performed by Rupali Khatun. This work was initially perfromed in the Electronics and Communication Sciences Unit (ECSU), Indian Statistical Institute (ISI), Kolkata in association with the Indian Unit for Pattern Recognition and Artificial Intelligence (IUPRAI) at Indian Statistical Institue (ISI), under the supervision of Mr. Angshuman Paul (Senior Research Fellow, Electronics and Communication Sciences Unit, Indian Statistical Institute, Kolkata) and Dr. Abhishek Roy (Former H.O.D, Amity Institute of Information Technology (AIIT), Amity University, Kolkata). Later on, this was updated, finalized and presented at the 2018 IEEE 8th International Advance Computing Conference (IACC 2018) under the supervision of Mr. Soumick Chatterjee (Research Fellow, Otto-von-Guericke University, Magdeburg, Germany)
The paper has been published with DOI: 10.1109/IADCC.2018.8692135 Preprint available at : https://arxiv.org/abs/1905.08611