Magnetic Resonance Imaging (MRI) has been widely used for both clinical diagnostics and in-vivo research. Although MRI images are acquired in the complex number field, traditionally magnitude images are utilized for most applications, while the phase of the complex numbers is ignored due to various signal processing and visualization issues. However, there is substantial evidence that a phase of MRI image can contain additional useful diagnostic information. For example the phase of MRI image can be used to detect excessive iron accumulation associated with many neurodegenerative disorders such as A1zeimer's and Parkinson disease. Phase wrapping and background phase variations are two main problems, which prevent clinical use of phase images directly. In this paper, we propose 2-D technique for phase unwrapping and background phase correction based on spectral image segmentation and detrending algorithms. This method handles various noise levels successfully, and most importantly, its extension to 3-D volumetric MRI phase image processing is conceptually straightforward.