MR image enhancement for ICH classification

Intracerebral hemorrhage (ICH) and cerebral amyloid angiopathy (CAA) are two common causes of hemorrhagic stroke. Distinguishing CAA from ICH is challenging due to similarities in clinical presentation and imaging. Magnetic Resonance Imaging (MRI) is one imaging modality that can be used in the early detection of ICH. The objective of this paper is to identify the best enhancement techniques in MR images for ICH classification. 5 samples of MR images with a final diagnosis of primary ICH (pICH) were selected from UKM Medical Centre (UKMMC). The MR images which were in Digital Imaging and Communications in Medicine (DICOM) format were later processed using the Matrix Laboratory (MATLAB) software. In this study 5 MR images from 5 different patients were processed using 4 image enhancement techniques; power law transformation histogram equalization image sharpening and median filter. All the 5 processed images were compared to obtain the best enhanced images based on their Absolute Mean Brightness Error (AMBE) and entropy values. Median filter is the best image enhancement technique with average values of AMBE and entropy of 0.0885 and 5.1472.