Computer Aided Diagnosis For Stroke Patients Based On Difussion-Weighted Magnetic Resonance Images

Mohd Noor, Nor Shahirah (2016) Computer Aided Diagnosis For Stroke Patients Based On Difussion-Weighted Magnetic Resonance Images. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Computer Aided Diagnosis For Stroke Patients Based On Difussion-Weighted Magnetic Resonance Images.pdf - Submitted Version Download (513kB)

Abstract

Stroke is a disease that occurs when a blood clot blocks an artery or a blood vessel breaks, interrupting blood flow to an area of the brain. Diffusion-weighted magnetic resonance imaging is increasingly having an important role in the diagnosis of stroke diseases. This medical imaging technique provides higher pathologic or lesion contrast for early stroke detection based on diffusion of water molecules in brain tissues. Key diagnosis and treatment planning is obtained from DWI, in order to assess early signs of acute stroke and to rule out hemorrhage. The major trouble with medical segmentation is the accurate segmentation in improving the treatment and diagnosis of disease due to use of medical imaging techniques. Image segmentation becomes challenging and complex task because of the usual medical image has unknown noise and inhomogeneity. The brain image segmentation is a complicated and challenging task that its precise segmentation is extremely important for detecting stroke. Within this context, this study proposes a technique for automatically detect stroke lesions of acute stroke and chronic stroke. An analytical framework of the stroke lesions consists of three stages which are pre-processing, segmentation, and performance evaluation. For segmentation process, fuzzy C-means integrated with correlation template are proposed to segment the lesion region. The algorithm performance was then evaluated using Jaccard index and Dice index. The results are 0.75 and 0.52 for average Jaccard index acute stroke and chronic stroke respectively.The average Dice index acute stroke and chronic stroke is 0.84 and 0.53 respectively. This method can be used to segment the lesions for acute stroke precisely but not too accurate for chronic stroke.

Actions (login required)

View Item