New Features Extraction Based on MRI Brain White Matter and Small Vessel Stroke Predisposition for Neural Network Input Classification

This Magnetic resonance imaging (MRI) is a very effective yet non-invasive medical imaging technique for clinical diagnosis and monitoring the abnormalities in neurological disorder. This paper provides a summary of current imaging and processing technique on MRI. Also includes in the review is the clinical features extract from MRI images for neural network classification system input. This review is focusing on white matter (WM) of brain since it has higher correlation to small vessel stroke occurrence. In other word, the assessment of white matter disease may be valuable in predicting future risk of stroke. Hence the proposed work for this study is focusing on WM features extraction from MRI images by using image processing technique includes noise removal or filtering. In medical image processing, poor image quality will result in poor feature extraction outcome which may lead to non-effective analysis, recognition and quantitative measurements. Therefore, pre-processing steps: i.e. noise elimination is a must for medical images processing as well as image segmentation. All the outcomes from image processing technique will be proposed to serve as attributes for classifier networks so that in future the classification performance can be evaluated for its accuracy, sensitivity and specificity.