Research On Segmentation Of Femoral Head Based On Levelset

Necrosis of the FemoralHead continues to be a significant public health problem in the World. At present, Early detection and therapy is the key to improving Necrosis of the FemoralHead prognosis at present. Currently, the supplemental use of MRI for Necrosis of the FemoralHead is an area of active research, and has proven to be a valuable adjunct to mammography in the detection and classification of the disease. Being an assistant to doctors, Computer-aided diagnosis applied to early Necrosis of the FemoralHead detection is a novel method in the disease treatment field, which greatly improvesing image qualities as well as being an assistant to doctors andbrings acquire diagnosis enhancement.Image segmentation is the most difficult task in image processing. It's a key point critical and essential component of image analysisin computer-aided diagnosis, and pattern recognition system, is one of the most difficult tasks in image processing, and determines the quality of the final result of analysis. If we segment the different FemoralHead pictures exactly, we can compare the area of the ischemia with the area of skeleton to decide the degree of the the patients disease. Because there are two perspectives pictures, we can approximately calculation the volume of the ischemia inside FemoralHead. So we segment the pictures exactly, it can get good prepare for the later auxiliary diagnosis of which period of the necrosis of the FemoralHead.This paper analyzed and compared several ultrasound image segmentation methods in Necrosis of the FemoralHead and made conclusions for their advantages and disadvantages.Firstly propose a new method based on a levelset model based on compound of region information and edge information.: this model have two parts, one be used to handle region information, use the energy function be established by the method of maximum likelihood function, another one be used to handle edge information, adopting Geodesic Active Contour model. Both region and edge messages are considered, which can not be achieved by previous algorithms. Getting a good result from different assessment experiment.Secondly propose add a punishing item to the levelset model, which restrict it in distance function. Reducing the time consume a lot.Finally propose a enhance algorithm in fuzzy field. Transform gray field to fuzzy Field and enhance in fuzzy field, then inverse transform. Getting a good enhanced result.