Dong Ying, Ji Jim
Department of Electrical and Computer Engineering, Texas A&M University, College Station, TX 77843-3128, USA.
Annu Int Conf IEEE Eng Med Biol Soc. 2010;2010:3309-12. doi: 10.1109/IEMBS.2010.5627494.
Phase unwrapping is a classical problem in Magnetic Resonance Imaging (MRI), Interferometric Synthetic Aperture Radar and Sonar (InSAR/InSAS), fringe pattern analysis, and spectroscopy. Although many methods have been proposed to address this problem, robust and effective phase unwrapping remains a challenge. This paper presents a novel phase unwrapping method using a region-based Markov Random Field (MRF) model. Specifically, the phase image is segmented into regions within which the phase is not wrapped. Then, the phase image is unwrapped between different regions using an improved Highest Confidence First (HCF) algorithm to optimize the MRF model. The proposed method has desirable theoretical properties as well as an efficient implementation. Simulations and experimental results on MRI images show that the proposed method provides similar or improved phase unwrapping than Phase Unwrapping MAx-flow/min-cut (PUMA) method and ZpM method.
相位展开是磁共振成像（MRI）、干涉合成孔径雷达和声纳（InSAR/InSAS）、条纹图案分析以及光谱学中的一个经典问题。尽管已经提出了许多方法来解决这个问题，但稳健且有效的相位展开仍然是一个挑战。本文提出了一种使用基于区域的马尔可夫随机场（MRF）模型的新型相位展开方法。具体而言，将相位图像分割成相位未包裹的区域。然后，使用改进的最高置信度优先（HCF）算法在不同区域之间展开相位图像，以优化MRF模型。所提出的方法具有理想的理论特性以及高效的实现方式。在MRI图像上的模拟和实验结果表明，所提出的方法与相位展开最大流/最小割（PUMA）方法和ZpM方法相比，提供了相似或更好的相位展开效果。
