A Crossbred Approach for Effective Brain Stroke Lesion Segmentation

Abstract

An Ischemic stroke is expressed as lost neurological brain work because of the sudden loss of blood dissemination in the specific territory of the brain. The sub-acute ischemic stroke is the most basic illnesses reason for death on the planet. In this paper we utilize a hybrid way to deal with detecting the ischemic stroke from the alternate pathologies in magnetic resonance (MR) images utilizing Kernelized Fuzzy C-means (KFCM) clustering with adaptive threshold algorithm and the Support Vector Machine (SVM) classifier. In the existing method, the Otsu's method incorporated with SVM classifier method is utilized for the segmentation of the ischemic stroke image, but it has the limited accuracy 88%, specificity 66% and the sensitivity value is 94%. For the exact identification and segmentation the KFCM algorithm is utilized. The distance and the intensity of the lesion tissue is identified by this method. The accuracy and segmentation aftereffects of the classifier are measured in the testing and training phase by looking at the comparable and a decent variety of sample sets by considering diverse groupings. Our test comes about demonstrating that, the performance of the proposed technique is assessed in view of the precision, recall, sensitivity, accuracy and overlap metrics of the framework. Compared with the existing classification method, the proposed method has 17.64% RMSE (Root Mean Square Error), 6.24% MAPE (Mean Absolute Percentage Error) and 2.55% MBE (Mean Bias Error), consumption time is 6.45 (s) and also the sensitivity and the accuracy ranges are 98.8% and 99%. The proposed approach is actualized using MATLAB and the realtime datasets are used for our examination.

Deep Learning Framework to Detect Ischemic Stroke Lesion in Brain MRI Slices of Flair/DW/T1 Modalities

“…The best result achieved with the proposed work was then compared with the existing methods, and the outcome is presented in Figure 14. The earlier works presented in the literature considered only a single MRI modality for the examination, and hence, the previous results show improved accuracy in [38][39][40]. However, compared with the works in [35][36][37], the results achieved with the proposed technique are better for the Flair and DW modalities.…”

MRI Images, Brain Lesions and Deep Learning

“…There are several preprocessing steps necessary to reduce the noise and artifacts in the medical images, before the segmentation [34,160,161].…”

“…The preprocessing steps commonly used are (1) the grayscale conversion, and the image resizing [161] to get better contrast and enhancement, (2) bias field correction to correct the intensity inhomogeneity [24,160], (3) image registration, a process for spatial alignment [160], and (4) removal of nonbrain tissue such as fat, skull, or neck which have intensities overlapping with intensities of brain tissues [21,160,162].…”

“…[32,66,180] report values of DSC between 0.80 and 0.86. Melingi and Vivekanand [161] reported that through combination of the Kernelized Fuzzy C-Means clustering and SVM they achieved an accuracy of 98.8% and sensitivity of 99%.…”

“…In the table 4 is noted that even though there are a large number of publications related to stroke ischemia (27 documents) and most deal with classification of stroke patients versus normal controls, or prediction of post-stroke functional impairment or treatment outcome [15,19,20,22,27,32,36,59,60,65,67,68,131,132,136,143,[148][149][150][151][152][153][154][155][156][157][158], there is a paucity of results related to demyelinating disease alone. However there are some publications dealing with Multiple Sclerosis (MS) which is the most common demyelinating disease (2) [159,160].…”