Segmentation Of Carotid Artery Plaque Base On Multi-MRI Image

Strong the subjectivity limitations of manual processing images, a waste of time and energy and other reasons as well as the rapid development of computer image processing technology, computer-aided medical diagnosis has become an inevitable trend. Organizations interested in composition from the large amount of data in the medical image data and analysis is an urgent need to solve the research topic. Among them, the image segmentation of both the initial critical step is also key to ensure that the subsequent processing operations correctness. Medical image processing, faced with the problem of image quality, the challenge of individual differences in human tissues and time deformation. Therefore, the introduction of a priori knowledge to commence for carotid MRI medical image segmentation method, the main work and the results are as follows:1. The image pre-processing process, from registration, sharpening, filtering, standardization, expand and improve image quality, reduce noise signal to enhance the texture, and other useful information imaging results, do a good job paving the way to prepare for the next step of segmentation. Then, combined with the pseudo-color technique, in four sequences, rapid interactive feature point registration.2. A wall segmentation results as a form a priori, to build a priori model of an approximate ring to achieve energy minimized to obtain the contour of the outer wall. Segmentation algorithm on the basis of the outline to determine, using the Bayesian classifier, a joint multi-channel image, the use of the existing limited data, the introduction of additional features, to identify plaque composition in the lumen of a framework to measure plaque characteristics.3. Apply SSVM algorithm to the segmentation of the plaque, and the results were compared, showed that in more image noise, of SSVM. Better than the former. Exclude misclassification phenomenon of noise, but also has some shortcomings. For small plaque composition, there will be able to distinguish the case.Through the work of more than two fully explain the more complex shape to be split, the introduction of a priori knowledge will reduce the segmentation difficulty, and accordingly enhance the accuracy of the segmentation. The introduction of structural information will increase the smoothness of segmentation.