Quantitative image quality evaluation of fast magnetic resonance imaging

Magnetic resonance imaging (MRI) has been widely used in the diagnosis and treatment of various diseases. The principle of encoding object contrast in the resonance spectrum gives MRI almost unlimited ways to acquire and reconstruct an image, but at the same time, arises the problem of how to effectively evaluate and optimize these parameters and methods. The perceptual difference model (CASE-PDM) is a quantitative image quality evaluation tool developed in our lab and has been successfully applied to several applications of MRI. It determines the visual difference between a high quality reference image and a more quickly obtained, but possibly degraded, MR image. A low CASE-PDM score indicates similarity of the images, and there is a threshold, below which, one cannot detect a difference between the two. The goal of our research is to use the CASE-PDM model to evaluate the image quality in fast MRI and to optimize the parameters. Two Human subject experiments, DSCQS and ADF, were designed and performed to validate the CASE-PDM model. The CASE-PDM predictions are closely correlated with the human subjects predictions, for both low-degradation and high-degradation images. We also found CASE-PDM threshold values that correspond to a "non-perceptible" difference. Using the above validation results, we focused on important variables in fast MR imaging, especially in spiral MRI and advanced parallel MRI techniques, including the SENSE regularization techniques and the GRAPPA reconstruction techniques. Data was generated from real scanners or simulated from the digital phantoms. Various acquisition parameters and reconstruction algorithms were combined to generate thousands of images for each application. CASE-PDM was used to quantitatively compare those images, and the CASE-PDM scores was analyzed statistically to give helpful hints for the engineers. In addition, a new modification of GRAPPA method called Robust GRAPPA was proposed. Robust GRAPPA was compared with conventional GRAPPA and obvious improvement in image quality was observed.