Abstract
Phase-sensitive inversion-recovery (PSIR) imaging may provide enhanced T(1) contrast. However, clinical implementation of PSIR imaging is hindered because image reconstruction with this method often lacks robustness and requires manual intervention, particularly for data acquired in multiple slices and with phased-array coils. In this paper, a new algorithm suitable for automatic PSIR image reconstruction of multislice and multicoil data is presented. This algorithm phase corrects by region-growing, employing both the magnitude and the phase information of image pixels. Specifically, phase gradients of the original complex image are first calculated and then used to determine the sequence of the region-growing. The signal direction relating to the phase error for each pixel is then determined during the region-growing using both the magnitude and the phase of the previously determined pixels that are located within a boxcar neighborhood of the pixel. Finally, the intrinsic intercoil and interslice correlation is exploited to ensure consistency in the global polarity of all of the PSIR images. The results are demonstrated with in vivo human brain images acquired at 3 Tesla with an eight-channel phased-array coil.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call