Cerebral Microbleed Detection by Extracting Area and Number from Susceptibility Weighted Imagery Using Convolutional Neural Network

Abstract

Cerebral microbleed (CMB) the small vessels in the brain which is one of the major factors used to facilitate in the early stage diagnosis for Alzheimer’s disease detection. In traditional, CMBs detection can be done manually by the neurologists, doctors or specialists. However, the process is time-consuming and the results are not accurate depending on the doctor experiences. Therefore the efficient and reliable of the automatic detection of CMB is needed. This paper proposes a new framework for CMB detection which employs segmentation of the region of interests (ROIs), detection of the CMBs and identification of the area from SWI scan images. Convolutional Neural Network(CNN) is applied to generate the desired models for later prediction. Shape matching mechanism is also applied to identify locations of CMB in the brain. The experimental result shows that the CMB can be classified with a recorded accuracy value of 95.45%. The CMBs were discovered from three different locations include (i) cortical region, (ii) cerebellum and (iii) brainstem with an accuracy value of 100%.

Review of methods for automatic cerebral microbleeds detection

“…The transfer-learning idea is to use a pre-trained network that has already learned some image features and fine-tune it on the particular dataset. Several networks were used for this purpose, including: AlexNet [131] in [126], ResNet50 [15] in [14], Faster-RCNN [129] in [18], VGG [132] in [125], U-Net [63] in [6], YOLOv2 [130] in [1,30], DenseNet 201 [133] in [4] or SSD [62] in [19] with the modification of feature enhancement. Sometimes, especially in case of SNP algorithm usage the detection task was substituted by classification of small fragments of image using either CNN or ResNet50 for instance in [14,77].…”

“…Although their reliability based on intra-and inter-observer agreement is reported, details of the methods used are usually not described [109]. [32,33], Traumatic Brain Injury (TBI) [45,29,5,44,140], stroke [31,5,73,20], Intracerebral Haemorrhages (ICH) [34,20], gliomas [26,51,17], hemodialysis cases [5], Cerebral Amyloid Angiopathy (CAA) [34], atherosclerosis [6], or did not distinguish any particular disease besides the appearance of CMBs [1,30,42,43,3,80,79,24,18,19,76,22,13,72,20,126,138,137]. Datasets used in the first category of researches focused on AD [81,82,83,36,84,85], SMART [37], TBI [86], stroke [86,…”

“…Further, there are typically neurological masks such as cerebrovascular fluid (CSF) mask, gray-white matter (GWM) mask and white-matter (WM) mask. Masks were generated in [32,73,33,35,26,41,28,27,45,29,78,17,126,23,76,2,72].…”

Deep learning-assisted IoMT framework for cerebral microbleed detection

“…They achieved 98.869 % sensitivity and 97.681 % accuracy in their CMB detection. Sa-ngiem et al [ 53 ] proposed CMB detection using area of interest-based segmentation (ROI), identifying the area of the CMBs from the images of SWI. This study employed the mechanisms of shape matching to locate CMBs in the brain MRI, achieving 95.45 % accuracy.…”