| Partial volume effect is a common problem of variety medical imaging equipment. The measured value is falling into the average of these matters when there are two or more different density matters in the same scan level. This phenomenon is called Partial Volume Effect.Positron emission tomography is the more advanced clinical examination image technique in nuclear medicine, with high sensitivity, high specificity, able to take image of whole body in one time, good security compared to other image technique. PET can location the lesion by tracer, when lesion at the molecular level. The partial volume effect of PET is more significant than CT/MRI due to imaging theory and detection technical constraints. Every point of lesions expands into a distribution by a function. Partial volume effect makes the image distortion, reduce the radiation concentration, enlarge the lesions size and blur the edge information. It is a main factor to impair the image quality and the accuracy of the quantitative analysis of PET. Partial volume correction is a difficult task in nuclear medicine.The current partial volume correction method of PET contains:1,recovery coefficient correction method, which is the simplest and most commonly used method. Recovery coefficient equal to radioactivity concentration measured on the observed image and the real radioactivity concentration. The value is affected by the lesion size and the size of region of interest. This method is difficult to apply in clinical because the lesion size is hard to define.2,merge with tracer kinetic model. This method use physiological parameters to estimate the partial volume effect when merge it into tracer kinetic model.3,method based on MRI/CT image, which use MR anatomical image for lesion location, and design algorithm to restore the real activity concentration. This method mainly used to study the PET brain image and required a high-resolution CT or MRI anatomic images to accurate registration with PET image. 4,method based on image theory and optimization theory. This method need to establish a model and is complex and time-consuming.Magnetic Resonance Imaging is a non-invasive approach of mapping the internal structure of the human body, which is reconstruction by the signal resonance of water nucleus in a magnetic field. Compared with other modalities of medical image, MRI has many advantages, such as with no ionizing radiation, imaging with multiple parameters in arbitrary plane, and providing excellent soft-tissue contrast. These advantages make MRI a very important tool for both research and clinical applications.Along with its development, MRI always encountered the problem of artifacts, such as metak chemical shift,motion,truncation and partial volume effect artifact. We focus on the partial volume effect in image processing artifacts. Partial volume effect brings in false information and fuzzy details in the image, affect the quality of MR imaging, which affect the accuracy of clinical diagnosis. Therefore, its correction is essential.The current partial volume correction method of MRI contains:1,image interpolation method, include nearest neighbor, bilinear, bicubic interpolation and so on. This method use only gray scale information and can bring in blurry result.2,method based on non-linear diffusion, which filter the homogenous regions and enhance the edge by controlling the diffusion after interpolated usually. This kind of method can achieve impressive visual quality enhancement but may generate extra artifacts.3,classification method, which establish the image histogram model, then estimate and classify the mix tissue in each voxel use spatial information between voxels. The results depend upon the accuracy of modeling and often need to specific parameters.Anisotropic diffusion is selective smoothing process which can achieve smooth the noise while maintain the characters of image boundary. Reverse diffusion is the implementation of diffusion equation inverted, which gray scales diffusion from low area to high. Partial volume effect is an artifact that blurred the boundaries between two matters, which can be corrected by reverse diffusion.In this paper, we analysis and research the causes and correction methods of partial volume effect.A PET partial volume correction method is proposed based on MRI image, which use MR anatomical image for lesion location, and design algorithm to restore the real activity concentration. Then analyze the changes of cerebral blood flow with age by statistical analysis the corrected data.We proposed an MRI partial volume correction method based on inverse diffusion according the anisotropic reverse diffusion method can achieve both smooth image and enhance the edge. This method is directly applied to MRI images, which can achieve the reduction of artifacts while enhancing the edge by the diffusion of pixel gray value. Experimental results show that our method works effectively well.We proposed the inverse diffusion method to correct the Gibbs artifact. Experiments show that the Gibbs artifact is effectively reduced and the valuable details of image are maintained.We developed a further follow-up study on blood vessels segmentation. The gray scales of blood vessels in MRA images suffer a great range of variety, especially the small blood vessels easily to mix with the surrounding tissue, make the segmentation of blood vessels difficult. In the paper, based on vessel's characteristics in MRA, We proposed a reverse diffusion method to intensify them, to ensure that the segmentation results are more accurate. Then we proposed a fast segmentation algorithm based on C-V model. |
PDF Full Text Request