Measurement of Signal-to Noise Ratio in clinical MR images

—Development of correction method of SD map using the identical ROI method—

Abstract

Signal-to-Noise ratio (SNR) is an important quantity factor used to describe the performance of a magnetic resonance (MR) imaging. However, it is difficult to measure accurately the SNR in clinical parallel MR images. This study is to develop a SNR measurement method using a modified identical region of interest (ROI) technique in clinical parallel MR images. Phantom experiments were performed on a 1.5 and 3.0 Tesla MRI machine to compare three different methods of SNR measurements. We also performed SNR measurements in clinical cases. The method is consisted of the following four steps: 1) Calculation of standard deviation (SD) map 2) Edge detection and removal on SD map 3) Interpolation of the removed edge region 4) Calculation of SNR map using the interpolated SD map. Our developed method were approximately consistent with those of other SNR measurement methods in both phantom and clinical images. In addition, the results of our developed method could measure the SNR in the edge region in parallel MR images. Our method makes it possible to simply and accurately analyze SNR of the clinical parallel image with a single scan.