Study And Application Of Supper Image Segmentation Methods In Medical Image

The data column of an image has been increasing drastically with the rapid development of information technology and digital imaging technology, and the abroad application of high resolution imaging equipment. Because of rich data, the data column has increased from many M bits and dozens of M bits to hundreds of M bits, even several G bits and dozens of G bits. Image segmentation is the basis of image analysis and image understanding. The traditional segmentation methods are not applied to supper image directly. So it is a new challenge for the supper image segmentation.This paper introduces a segmentation method for supper images. Firstly, we block the supper image whose block images are the same size, and number block images. Then block images orderly are inputted into memory, and each block image gray information is gotten. The supper gray information is gotten with the block images'gray information when all block images are processed. In this paper, the gray information is 2D gray histogram. And then the optimal threshold of the supper image is gained based on the 2D fuzzy entropy. Finally, block images of the supper image are segmented according to threshold segmentation, and the results are saved. Then the supper image is segmented.The resolution of medical image has been improved with the development of medical imaging technology. The resolution of medical microscope images which is gained by confocal laser scanning microscopy (CLSM) is 0.23μm/pixel or more. It means that there are about 4000 pixels per micron. So the resolution and data column is very high. Because these supper medical images have high resolution and abundant information, researchers can observe the detail of tissues'structure more accurately and deeply, which supply abundant information for diagnosis of disease and medical research. In this paper, the method of supper image segmentation is applied to the medical image. So we can get a higher level analysis and understanding for medical images.Experimental results show that the method presented in this paper has a good performance for supper images, and gets rid of the block effect. It is also anti-noise strongly.