Investigation and implementation of image processing algorithms for medical application

• Main Author: Mya Thandar Maung.
• Other Authors: Mohammed Yakoob Siyal
• Format: Final Year Project
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/54508
• Institution: Nanyang Technological University

Functional Magnetic Resonance Imaging (fMRI) has experienced a rapid growth in the past several years and has found further applications in a wide variety of fields, such as neuroscience, psychology, and political science, in addition to medical applications. Currently, there exist a number of different imaging modalities that allows the users to study the physiological changes that accompany brain activation. Each of these techniques provides a unique perspective on brain function and meeting the individual purposes. Among them, there has been a growing number of neuro-imaging studies performed using fMRI. The fMRI is now solidly established as a noninvasive diagnostic technique in acquisition of physiological and biochemical information and in particular for studying brain activity, as well as tumors and cancerous vicinity. The fMRI takes advantage of the relationship of certain stimuli leads to initiate changes in neuronal activity, which give momentary changes in blood oxygenation/oxygen level (BOLD) to the active region of the brain. In practice, it identifies the brain activity by picking up minute changes in blood flow in response to stimuli that the subject or patient experiencing associated with alternative non-stimulated instance or pause while the scanning is in progress. Changes in the measured signal between individual images are to make inferences regarding task-related activations in the brain. This paper discuses the analysis of fMRI data, from the initial raw data to its use in locating brain activity and to map the brain function. A standard fMRI study gives rise to massive amounts of noisy data as its signal is corrupted by random noise and various components that arise due to both the system hardware reasons and the subjects themselves together with a complicated spatial-temporal correlation structure. This is further denoising those nuisance signals and making corrective measures for the reduction of noises correspond to the subjects. The signal data then have to undergo statistical processing for the development of image representation. The Statistics plays a crucial role in understanding the nature of the data and obtaining relevant results that can be used as diagnostic tools for medical practitioners and as evidences to be interpreted by scientists.