Algorithms – MICLab

Extended 2D consensus hippocampus segmentation

MIC-Suporte — Sat, 09 Apr 2022 12:32:19 +0000

Extended 2D Consensus Hippocampus Segmentation

Medical Imaging with Deep Learning 2019

Hippocampus segmentation plays a key role in diagnosing various brain disorders. Nowadays, segmentation is a manual, time consuming task and considered to be the gold-standard when evaluating automated methods. For years the best performing automatic methods were multi atlas based with 80 to 85% DICE and time consuming, but machine learning methods are recently rising with promising time and accuracy performance. In thiswork, a novel method for hippocampus segmentation is presented, based on the consensus of tri-planar U-Net inspired CNNs, with some modifications based on successful CNNs of the literature, and a patch extraction technique employing data from neighbor patches. Our in-house dataset has hippocampus atrophies resulted from epilepsy surgery treatment. Our method (labeled e2dhipseg) achieves cutting edge performance of 96% DICE in ourtest data. Our method was also compared to other recent methods in the public ADNI and HARP datasets.

Full paper here: https://openreview.net/pdf?id=Sygx97DaKV

iamxt: Max-tree toolbox for image processing and analysis

MIC-Suporte — Thu, 07 Apr 2022 12:29:37 +0000

iamxt: Max-tree toolbox for image processing and analysis

SoftwareX, Volume 6, 2017, Pages 81-84

The iamxt is an array-based max-tree toolbox implemented in Python using the NumPy library for array processing. It has state of the art methods for building and processing the max-tree, and a large set of visualization tools that allow to view the tree and the contents of its nodes. The array-based programming style and max-tree representation used in the toolbox make it simple to use. The intended audience of this toolbox includes mathematical morphology students and researchers that want to develop research in the field and image processing researchers that need a toolbox simple to use and easy to integrate in their applications.

Full paper here: https://doi.org/10.1016/j.softx.2017.03.001

Divergence map from diffusion tensor imaging: Concepts and application to corpus callosum

MIC-Suporte — Mon, 04 Apr 2022 12:11:43 +0000

Divergence map from diffusion tensor imaging: Concepts and application to corpus callosum

Engineering in Medicine and Biology Society (EMBC), 2016

This work proposes a novel approach to the analysis of Diffusion Tensor Imaging (DTI) by applying the mathematical concept of divergence, used in vector analysis. This is achieved by choosing an arbitrary direction of analysis and using this direction to transform the diffusion tensor field into an oriented vector field. The method was inspired by the idea of imposing a liquid flow inside the biological tissues, oriented in the direction of analysis, and watching the direction it would be expected to take as it flows through the paths created by the fibers. The experiments were conducted for the particular case of the analysis of the corpus callosum, using real DTI from several subjects. Results showed that the divergence map allows extraction of useful information about the spatial organization of the corpus callosum, providing a way to determine a reference plane that could be used, for example, in studies involving intersubject comparison.

Full paper here: https://doi.org/10.1109/EMBC.2016.7590900

A comparison between k-Optimum Path Forest and k-Nearest Neighbors supervised classifiers

MIC-Suporte — Sat, 02 Apr 2022 12:00:26 +0000

A comparison between k-Optimum Path Forest and k-Nearest Neighbors supervised classifiers

Pattern Recognition Letters, Volume 39, 1 April 2014, Pages 2-10

This paper presents the k-Optimum Path Forest (k-OPF) supervised classifier, which is a natural extension of the OPF classifier. k-OPF is compared to the k-Nearest Neighbors (k-NN), Support Vector Machine (SVM) and Decision Tree (DT) classifiers, and we see that k-OPF and k-NN have many similarities. This work shows that the k-OPF is equivalent to the k-NN classifier when all training samples are used as pro- totypes. Simulations comparing the accuracy results, the decision boundaries and the processing time of the classifiers are presented to experimentally validate our hypothesis. Also, we prove that OPF using the max cost function and the NN supervised classifiers have the same theoretical error bounds.

Full paper here: https://doi.org/10.1016/j.patrec.2013.08.030