Magnetic resonance phase unwrapping method based on pixel classification and local surface fitting

Abstract

The invention discloses a magnetic resonance phase unwrapping method based on pixel classification and local surface fitting. The magnetic resonance phase unwrapping method comprises the steps of: calculating a phase local deviation map of a complex vector with an amplitude being 1 and a phase being true phase, and then calculating a phase local deviation map of a wrapped phase; dividing pixels in a region of interest into two types including disconnected blocks and residual pixels; identifying whether any of the disconnected blocks contains the wrapped phase, if so, performing in-block phase unwrapping on the disconnected block, wherein the in-block phase unwrapping is implemented by dividing the disconnected block into two types including disconnected subblocks and residual pixels, conducting phase unwrapping among the disconnected subblocks preferentially, and then conducting phase unwrapping among the residual pixels; and finally carrying out in-mask phase unwrapping, wherein the in-mask phase unwrapping is implemented by combining the disconnected blocks in the region of interest, and then conducting residual pixel unwrapping in the region of interest by utilizing the unwrapped disconnected blocks. The magnetic resonance phase unwrapping method can be used for extracting the pixels with fast-varying true phases and the pixels with fast-varying wrapped phases in the region of interest more accurately, thus can realize phase unwrapping more reliably.

Description

technical field [0001] The invention relates to the technical field of magnetic resonance phase unwrapping, in particular to a magnetic resonance phase unwrapping method based on pixel classification and local surface fitting. Background technique [0002] Magnetic resonance imaging (MRI), due to its advantages of no ionizing radiation, rich tissue contrast information and non-invasive detection, has been widely used in clinical medical imaging examinations. One acquires MRI data as complex numbers, consisting of magnitude and phase. Typically, magnitude images are heavily used, while phase information is discarded. However, phase information has some very important clinical application requirements, such as chemical shift imaging, susceptibility imaging, blood flow velocity measurement in magnetic resonance angiography imaging, human brain phase imaging, etc. However, obtaining real phase images is not an easy task. This is because the obtained phase is only defined in t...

AI Technical Summary

Problems solved by technology

However, when encountering severe noise pollution or 2π jumps, suboptimal solutions may be obtained, and residual windings appear in the resulting image

Images