Liver Ultrasound Images Boundary Extraction And Assistant Interpretation

The technology of medical image processing is an important part for the setting up PACS system. Ultrasound diagnosis is an important medical image diagnosis means like X-ray, CT, isotope scanning and MRI. The diagnosis about liver on the base of ultrasound image is the important method. But, compared with other medical images (e.g. CT and MRI), the quality of ultrasound images is relatively low. At present, the diagnosis about liver is determined the nature, the conclusion of disease is influenced by doctors.An important approach for describing a region is to quantify its texture content. In this paper, we compute texture statistical parameters by analyzing features of the liver ultrasound image texture, which help doctors to work more efficient.Considering effect on the liver ultrasound images of the noises, we need pre-process images. In the paper, we introduce both smooth and sharpen methods. For the feature of the liver B-ultrasound image, we adopt the Gaussian low-pass filter and contrast-stretching transformation methods. To attain the exact edge, we research the edge identification features using energy active curve and geometry active curve, but the two methods are not perfect and do not use in clinic. Through a lot of examinations, at last, we select a simulated annealing algorithm that combines both methods to segmentation, and use it in the liver B ultrasound images boundary extraction. This algorithm is put forward by optimized iterative way, it not only enhances the whole constringency, but also quickens the speed of the evolution, then gets the optimized values.On the base of successful gaining the boundary detection features, we use the knowledge of the processing image, then compute the center of gravity coordinates, perimeter, area of the disease, which are the base of the diagnosis for the doctors.