Segmentation Of Ultrasound Images With Fuzzy Enhancement

Ultrasound image diagnosis is an important medical image diagnosis means like x-ray CT, isotope scanning and MRI. To facilitate manual or computer-aids analysis, in this paper, we make a fuzzy-enhancement on the ultrasound images as a preprocessor of the clinician diagnosing and the computes aids analyzing. However, ultrasound B-scan images often have low contrast and the characteristic speckle noise. These cause major problems for image analysis, both by manual and computer-aided techniques, particularly the computation of quantitative measurements.Image enhancement is one essential technique of the digital image processing techniques. It is an important link in the image processing system; more over, its result has direct relations with the high-level process and relations with understand of the image. This paper presents a new fuzzy enhancement methods of image enhancement based on analyze and summarize model and techniques in existence. As we know, some useful detail of the images may be lost by the normal image enhancement technique. It's hard to get the accurate histogram distribute what we really want in advance. The fuzzy set theory solves the complicated problems made by uncertain factors. In this new method, we introduce a processing control parameter. Changing the processing control parameter is good for the processing optimization and ultrasound image contrast enhancement.It is an important precondition and a base of ultrasound image diagnosis to effectively extract Region of Interesting from an ultrasound image. The disease can be diagnosed by analyzing changes of both the boundary and inside contents of the organic boundary. Image segmentation, partitioning an image into different regionswith some specific properties and labeling each pixel with its underlying class, has always been an important and challenging problem for many years. The main difficulties lie in the great variability of images and the presence of noises.Compared with other medical images (e.g. CT and MRI), ultrasound images are particularly difficult to segment the ultrasound image as the quality of the images is relatively low. So it is necessary to study the segment algorithm of ultrasound images based on the enhancement. There is a special kind of noise called speckle in ultrasonic images. The analysis both by manual and computer-aided techniques will be affected, especially the image segmentation based on threshold value. If there are a number of scatters in the body, the statistics of the speckle is Rayleigh distribution. Because there's speckle noise in the ultrasound images, it's hard to get the satisfied effect by segmentation based on threshold value. There're a lot of isolated spots in the result after segmentation. In this paper, Markov random field (MRF) and Gibbs random field are used as a prior model, in order that the complicated solution of the entire image information can be transformed to simply solution. It will be easier to deal with, and Bayesian segmentation method will be more practical. The spatial information among pixels is used to restrain the speckle noise and eliminate the isolated spots.In this paper, both image enhancement and image segmentation aspects are mainly studied, and present a new fuzzy algorithm to enhance the contrast of medical ultrasound images. Significant improvement is achieved in tissue contrast of the result images. We model the intensity of ultrasound image as gauss distribution based on the statistic of the speckle in the ultrasound image, and use Markov random field (MRF) and the maximum a posteriori theory to segmentation the enhancement images. The experiments on real medical ultrasound images prove that the proposed algorithm is insensitive to noise and shows the validity of the method of enhancement and segmentation and expectant experiment results are obtained.