The Research On Motion Artifacts In Medical CT

The motion artifacts in medical CT are that the reconstructed image is misty or unclear since the patient has the conscious or unconscious move in the process of scanning. In order to enhance the image readability and make the medical worker more accurate effective observation and diagnose for the human body anatomy structure and focus, evaluating and reducing the effect of motion artifacts has become a field that modern diagnose medical science takes more seriously.There are two very important indices in evaluating CT image quality, which are space resolution and density resolution. Space resolution and density resolution are called the high contrast resolution and low contrast resolution in medical field.High contrast resolution signifies the capability of distinguishing the minimal space interval in CT imaging. In general, space resolution is denoted by the minimal space interval of being able to differentiate. Low contrast resolution shows the ability of distinguishing the minimal density interval. We use the ratio of minimal density difference of being able to distinguish and the average density to express density resolution. In the processing of CT scanning, the movements which is caused by involuntary or uncooperative patient (it mostly happens in patients which are unable self-control or young) or the organ peristalsis in the scanning the visceral organs brings about the distortion of high contrast resolution and low contrast resolution, and then the motion artifacts affects the doctor's diagnose.For the third generation medical CT which is mostly used and based on the filter back projection algorithm, this paper explore and investigate the effects on accurate degree of high and low contrast resolution which is caused by the translation and rotation of patients. A method of evaluating concrete effects have been put up and verified by the experiment. Using this method, medical workers can discriminate the motion artifacts in the reconstructed image according to the experience value of organ's displacement in hand and manage to improve the accurate degree of diagnosing and treating.Going a step further, researchers have been exploring the new image reconstruction algorithm to restrain the formation of motion artifacts. The discussion is focus on the fields of accelerating scanning speed, improving the reconstructing algorithm and improving the data expressing precision. To get more accurate image by improving image reconstructing algorithm gets more attention by people among these methods.In many situations, we are not interested in the whole body but just some parts of body, so it is ok to reconstruct only the region of interest. But the traditional ray projection and reconstruction algorithm can't meet people's requirement owning to the global behavior. Thus some local reconstruction algorithms for the region of interest have been developed in the past few years.The traditional filter back projection (FBP) image reconstruction algorithm has the birth disadvantage which is seriously motion artifacts caused by longer scanning and reconstructing time. For local reconstruction algorithm only concern the region of interest, the required data of scanning reconstruction are reduced largely and the time of scanning reconstruction is also obviously shorten, have restrained the motion artifacts thereby.This paper mainly discusses three kinds local reconstruction algorithm. Lambda Tomography algorithm among them is a strict local reconstruction algorithm. The algorithm self is not limited in two-dimensional or three-dimensional field. And it can reconstruct any dimensional space information in mathematics theory. And that other two algorithms which are pseudolocal tomography and wave-based multi-resolution local tomography are not absolute strictly local, they need the projection data which is slightly bigger than ROI. Wave-based local reconstruction algorithm is still an active research direction at the present time.This paper has proposed a method of Helgason-Ludwig consistency condition (HLCC)-based and wavelet local reconstruction aiming at the characteristic of local reconstruction algorithm which could effectively reducing motion artifacts. The algorithm firstly extrapolates the sinogram data using HLCC, and then reconstructs the ROI using wavelet-based reconstruction algorithm. This paper has proposed an improved method to extrapolate the CT sonogram data, which uses an ellipse having six parameters. The six parameters are optimized using HLCC. Then we reconstruct the image using wavelet-based reconstruction algorithm. At the same time we have provide the project for dealing the uncontinuity between the measured data and extrapolated data. Furthermore, we have used different cost function to solve non-convergence problems and unsteady case. The method has decreased the scanning and reconstructing time and restrained motion artifacts, and has provided a new thought to develop image reconstruction algorithm for CT manufacturer.To sum up the above discuss, this paper is initiated and motivated by obstacle of diagnosis and treatment caused by motion artifacts in clinic. Two central issues are investigated in this paper. The first is that we have explored a algorithm of quantitative evaluating value on motion artifacts according to FBP-based medical CT which is mostly used at present. To meet the needs of future, the second part we have brought forward the HLCC-based wavelet local reconstruction algorithm. Since this algorithm has the locality behavior, it can effectively reduce the motion artifacts. The HLCC-based extrapolated method in the above algorithm has reduced largely the blurred border phenomenon which is often appeared in the wavelet-based local reconstruction algorithm.The contribution of this paper is summarized as follows,1. This paper presents the method of evaluating motion artifacts in the medical CT which is based on FBP algorithm. We have analyzed how the translation and rotation movement affect the high contrast resolution and low contrast resolution respectively and given the evaluating method. We set two adjacent ideal points and deduce the effects on high and low contrast resolution by analyzing the effects on point spread function (PSF) caused by translation and rotation movement. According to the integral formula of CT image reconstruction, we carry out analysis on the introduction and propagation of noise and provide the deducing process and conclusion. We have chosen some family circumstance in simulation experiment, Two adjacent ideal points and two circles which is close to each other are the experiment objects of study in the accurate degree of the high contrast resolution experiments. In the low contrast resolution experiments, we have chosen four pinholes with different diameters under homogeneous background as the object of study. All above experiment scenes are usually used in CT experiments. Our idea has been verified by the experimental result.2. Exploring to reduce motion artifacts is being always a field of which researchers who develop a new generation medical CT system pay great attention to. Among them, wavelet-based local reconstruction is the most active research field. Due to the truncation of local projection data, wavelet local reconstruction has also led to other false image especially in the border location of ROI while it effectively reduces the motion artifacts. This paper presents a local reconstruction algorithm which is based on HLCC and wavelet multi-resolution analysis. Firstly, we extrapolate the sinogram data using HLCC-based elliptical method, and then reconstruct local image by wavelet local reconstruction method. The experimental results prove the effectiveness of the algorithm in improving image quality. The border of ROI is better than results which use constant extrapolation sinogram data. This has provided real feasibility for the wavelet-based reconstruction algorithm which is focus on reducing motion artifacts.