The Correction And Realization Of MRI Motion Artifacts Based On Energy Constraint

With the development of medical technology, magnetic resonance imagingtechnology has become an indispensable advanced means of disease diagnosis.Because of its advantages of non-invasive, no radiation damage, high resolution andarbitrary direction faults, it has played a more and more important role in clinicaldiagnosis and scientific research. However, during the process of magnetic resonanceimaging, the patient’s movements were often difficult to avoid, resulting in MR imagereconstruction with the motion artifacts, which are affecting the doctor to make thecorrect clinical diagnosis at last. More and more scholars devote themselves toresearch on image processing algorithms. Because of the diversity of movement, themotion artifacts in MR image formation is complex. The purpose of this paper is tostudy the ideal rigid translational motion state of motion artifacts and artifactscorrection.Firstly, we describe the NMR principle and the principle of magnetic resonanceimaging. The basic theoretical knowledge of MRI and the generation mechanism ofMR motion artifacts are introduced.Then we study two recovery algorithms based on MRI energy constraints byusing simulation and imaging experiment, which are direction information measureand minimum entropy focus on criterion constraint.In the experiment, the direction information measure recovery method can’tcorrect the motion artifacts. It can’t be used as the image criteria, but this method canfind the boundaries in the image well. It can be used for other image processingalgorithms.In order to correct the motion artifacts properly, we use phase recovery algorithmmethod based on minimum entropy constraint and improve it. The key to realizationis to make the motion estimation accurately. We use optimal correction strategy,namely reversed motion simulation in k‐space frequency encoding direction and phaseencoding direction. We calculate the corresponding phase deviation and use theminimum entropy constraint in image domain. In the end, we use Fourier transformbefore reconstructing the image. We not only take improvement of the k-space linenumber in each iteration, but also change the original single step correction mode atthe same time. So we can shorten the imaging time and improve the accuracy ofcorrection. Experiments show that the translational motion artifacts in MRI can beeffectively suppressed by using the method.Finally, the motion artifacts correction platform is developed under the graphicuser interface in Matlab environment. The platform can display the simulation dataand experimental data, show the calibration results and the difference image. It is abasic platform for development of other algorithms.