A correction method for motion-induced artifacts in high b-value diffusion MRI

• Main Authors: Chien-Feng Huang, 黃芊丰
• Other Authors: Wen-Yih Isaac Tseng
• Format: Others
• Language: en_US
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/bdx6u4

碩士 === 國立臺灣大學 === 醫療器材與醫學影像研究所 === 105 === Diffusion MRI is becoming increasingly important for clinical and neuroscience studies owing to its capability to depict microstructural properties of white matter. For a more accurate estimation of diffusion index to reflect microstructural properties, recent advances in diffusion MRI techniques, such as diffusion spectrum imaging (DSI) and high angular resolution diffusion imaging, acquire diffusion-weighted (DW) images with multiple diffusion sensitivities and directions. Due to the use of strong diffusion gradients, these techniques are sensitive to head motion, which can result in signal dropout and image misalignment. These artifacts may then lead to errors in diffusion index calculation. Therefore, it is necessary to discard or correct these degraded data in the subsequent analysis. Currently, there are no effective methods that use post-processing algorithms to restore signal dropouts, in either clinical or high b-value diffusion MRI. As for misalignment, the most frequently adopted approach is to align the images to a reference image based on image features. Unfortunately, this method will fail to correct high-b-value diffusion MRI, owing to its blurred edges of the tissue and outline of the brain. Therefore, this thesis develops an automatic post-processing algorithm to correct signal dropout and misalignment for diffusion MRI. Images that would have been discarded can now be restored and made available for subsequent image analysis. This thesis is divided into four parts. First, we will explain the principle and algorithm for signal dropout detection. Second, we will explain the principle and algorithm for signal dropout correction. Third, we will explain the principle and algorithm for misalignment correction. Finally, we will combine the above three parts into an automatic motion correction algorithm and explain the details and steps of this algorithm. In summary, this thesis successfully develops a motion correction algorithm that can be used in high b-value diffusion MRI and can correct signal dropout and misalignment retrospectively. This algorithm can salvage precious images, such as images of patients with rare diseases, that would have been discarded previously and allow for subsequent analysis of these images. Furthermore, this method is easy to implement, and does not require any additional software and hardware.