Accelerated magnetic resonance imaging

• Main Author: Hong, Xinying.
• Other Authors: School of Electrical and Electronic Engineering
• Format: Final Year Project
• Language: English
• Published: 2011
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation::Medical electronics
• Online Access: http://hdl.handle.net/10356/45892
• Institution: Nanyang Technological University

Magnetic Resonance Imaging (MRI) is a great invention in the biomedical field, it is an instrument which is non-invasive, yet able to image the interior structure of the body and yield information which other techniques could not. This technique is widely used in diagnostic medicine; however it is also commonly used in the oil and food industries. This instrumental technique is safe to use medically as it uses magnetic field, which is proven to be of no harm to the human body. However, this MRI technique has been hampered by long scan time due to physical and psychological constraints. Therefore, Accelerated Magnetic Resonance Imaging is being explored to reduce this scan time which will in turn benefits the patients and economically wise, reduce costs. There are many ways to accelerate imaging speed, but due to time constraints, in this Final Year Project, we will only explore Compressed Sensing, which is also known as sparse sampling, and how MRI benefited from this reconstruction method. However, due to time constraint, only Cartesian CS trajectories are being focused on, although non-Cartesian CS has greater potential and more advantageous for some application. Three methods of CS Cartesian trajectories reconstruction are being explored in this paper by comparison with other linear reconstruction methods and CS reconstruction method is found to be able to produce better reconstruction images, where the applications requires rapid imaging and exhibit high resolution and high contrast image features. The most distinct artifact in CS is the loss of low-contrast features in the image. Compressed Sensing plays a major role in applications that are limited by scan time and when the images exhibits transform sparsity.