High-Resolution Imaging Using Keyhole-Based Compressed Sensing Averaging (CSAK) Technique

Abstract

This study is to propose a magnetic resonance (MR) imaging technique, called as CSAK (compressed sensing averaging combined with a keyhole), and prove its possibility through in vivo experiment for high-resolution and fast imaging. For CSA, undersampling ratios of 25 %, 35 %, and 45 % were compared with a full k-space. Each undersampling ratio was composed of three subsets. High resolution CSAK images from the phantom and human subjects showed the increased signal-to-noise ratio (SNR) and depicted the anatomical structures more clearly. In addition, images with a 25 % undersampling ratio showed the highest SNR and reduced acquisition time by approximately 30 % relative to that of a full image. This study suggests that the proposed undersampling method for images with higher spatial and/or temporal resolution could yield better results in medical applications. Therefore, the CS-averaging technique could solve the limitations of acquiring much higher resolution images in clinical 3T MR imaging.