Ischemic stroke classification using machine learning technique based on 3D MRI data

Nurul Ezati Aida, Amsah (2017) Ischemic stroke classification using machine learning technique based on 3D MRI data. Project Report. UTeM, Melaka,Malaysia. (Submitted)

Text (24 Pages) Ischemic stroke classification using machine learning technique based on 3D MRI data.pdf - Submitted Version Download (1MB)

Abstract

Magnetic resonance imaging (MRI) is a medical imaging technique that uses magnetic fields and radio waves to produce high-quality images of the body.It is a nonaggressive,no-radioactive and pain-free medical imaging system for visualizing and non-invasively detecting the stroke.An accurate automatic detection and classification of images is very important task for a proper medication because any delay or wrong diagnosis may become a fatal to the patient.Besides,an assessment of brain lesion in MRI is a complicated process and only can be performed by experienced neuro radiologists with significant degree of precision and accuracy.The result from the MRI scan only can be reviewed by the professional neuro radiologist and the task is time-consuming.The objective of this project is to design a technique for stroke detection and classification using Machine learning technique,to analyze brain MRI for stroke detection and classification and lastly,to evaluate the performance of the machine learning technique in the detection and classification stage.The Region of interest (ROI) that obtained from the segmentation stage will be analyzed for classification process.First order statistical approach is applied on the Region of interest (ROI) to extract the feature of MRI image and used as input to Support vector machine (SVM) classifier.It will show the characterization of the ROI of different type of ischemic stroke either acute or chronic lesion.After the classification stage,the performances evaluation of the system are verified.The performance of this classification system are accuracy,sensitivity and specificity.The results demonstrate that 100% accuracy has been achieved for both lesion.Last but not least,the Graphical User Interface(GUI) was developing to make the system user friendly and attractive.

Actions (login required)

View Item