Meeting Banner
Abstract #3460

Feature Extraction using Convolutional Networks for Identifying Carotid Artery Atherosclerosis Patients in a Heterogeneous Brain MR Dataset

Mariana Bento1, Luis A. Souto Maior Neto2, Marina Salluzzi3, Yunyan Zhang1, and Richard Frayne1

1Departments of Radiology and Clinical Neuroscience, Hotchkiss Brain Institute, University of Calgary, Calgary, AB, Canada, 2Biomedical Engineering, University of Calgary, Calgary, AB, Canada, 3Calgary Image Processing and Analysis Centre, Foothills Medical Centre, Calgary, AB, Canada

Analysis of pathology in patients from heterogeneous datasets using machine learning techniques provide valuable information for identifying patients with carotid artery atherosclerosis disease. We propose and evaluate a method to automatically identify these patients based only on MR brain imaging findings in a dataset also containing multiple sclerosis patients and healthy control subjects. The features extracted using convolutional networks were discriminative, showing high accuracy rates (>96%) to distinguish between the three classes: atherosclerosis patients, multiple sclerosis patients or healthy controls. The method may help specialists in the diagnosis (specially in critical cases), and evaluation of disease activity.

This abstract and the presentation materials are available to members only; a login is required.

Join Here