Automated segmentation and classification of brain stroke using expectation-maximization and random forest classifier

Magnetic resonance imaging (MRI) is effectively used for accurate diagnosis of acute ischemic stroke. This paper presents an automated method based on computer aided decision system to detect the ischemic stroke using diffusion-weighted image (DWI) sequence of MR images. The system consists of segmentation and classification of brain stroke into three types according to The Oxfordshire Community Stroke Project (OCSP) scheme. The stroke is mainly classified into partial anterior circulation syndrome (PACS), lacunar syndrome (LACS) and total anterior circulation stroke (TACS). The affected part of the brain due to stroke was segmented using expectation-maximization (EM) algorithm and the segmented region was then processed further with fractional-order Darwinian particle swarm optimization (FODPSO) technique in order to improve the detection accuracy. A total of 192 scan of MRI were considered for the evaluation. Different morphological and statistical features were extracted from the segmented lesions to form a feature set which was then classified with support vector machine (SVM) and random forest (RF) classifiers. The proposed system efficiently detected the stroke lesions with an accuracy of 93.4% using RF classifier, which was better than the results of the SVM classifier. Hence the proposed method can be used in decision-making process in the treatment of ischemic stroke. (c) 2019 Nalecz Institute of Biocybernetics and Biomedical Engineering of the Polish Academy of Sciences. Published by Elsevier B.V. All rights reserved.