On improving clustering algorithm in medical image segmentation.docVIP

 PAGE \* MERGEFORMAT 11 On improving clustering algorithm in medical image segmentation [Abstract] the characteristics of medical images, design a clustering algorithm for image segmentation and clustering genetic algorithm is introduced, the parallel genetic algorithm and random search of, starting from the DBSCAN algorithm, proposed for its limitations a sampling-based DBSCAN algorithm and genetic optimization to achieve good segmentation results. [Keywords:] medical image; clustering algorithm; genetic algorithm; split With the development of medical technology, medical diagnosis related to a variety of images in the modern disease diagnosis plays an important role in analyzing and reading grayscale medical images, image contrast, edge characteristics and SNR of the correct diagnosis of of crucial importance. but can not be avoided in the image shooting some noise (quantum noise and grain noise, CCD dark current noise, etc.) and pathological changes of small clear case of non-image information, the correct diagnosis of the disease, so in order to improve the rate of correct diagnosis of the disease, medical image processing technology is particularly important [1]. In recent years, medical image processing segmentation in the international field of image segmentation on a new research hot spot. In this method, the image map with a weighted undirected graph, the pixels as nodes. Image using the minimum cut criteria are the best segmentation, the method is essentially to image segmentation problem is transformed into an optimization problem. Is a point on the clustering method. For data clustering also has good prospects. This division medical diagnosis technology is very helpful. A clustering algorithm In recent years, large amounts of data are stored in spatial database, how to improve query efficiency and extract useful data from a large number of patterns is very important. Cluster analysis is a physical or abstract objects into a collec