A Study Of Skull Image Segmentation Based On Local Level Set Method

There are many types of diseases with skull and the clinical diagnosis has a certain complexity. The process of analysis of skull CT image by doctors is tedious and subjective. It needs to handle a large number of images and it is possible to have mistakes. In this case, automatic detection and diagnosis have been proposed. Through image processing technology, the sick area can be diagnosed by segmentation, feature extraction, reconstruction and recognition. And finally doctors identify them manually for further examination, which is more safe and effective to analyze the patient`s condition. In the fields of archeology and medical appraisal, researchers can also use image processing system to analyze skull image more convenient and efficient.Image processing technology has rapidly developed in recent decades. It has been widely applied in many fields such as medicine, military, meteorological and industry. Image segmentation is an important prime step in image processing. And it is also one of the fundamental procedures in the field of computer vision. There are many categories of image segmentation methods. Among them, Level Set method is a commonly used method for medical image segmentation and it belongs to methods based on geometric deformable models. Level Set Method is used to describe the evolution of the curve. Its biggest advantage is stability and topology-independent. Nowadays, Level Set method has been widely used in image processing, computer vision and robot navigation and other fields and it has a good prospect.This article discusses the principles of Level Set method and several improved methods based on Level Set. And it mainly studies the application of sparse field Level Set method and also proposes innovative local Level Set algorithm in skull CT image segmentation. Though some image preprocessing methods, the rough segmented area can be got as seed image and it can be used in iterated operation. And the paper demonstrates its superiority in the experiment and good applicability to get main parts of the skull CT image. In this paper, EM(Expectation Maximization) algorithm is also used to divide and isolate parts of the bone and soft tissue and the cavity section. It has certain significance for medical research and computer aided judgment and archeology.Through experimental results, it shows that the proposed improved local Level Set Method has good segmentation results and it achieves the desired targets, and it is also faster in computing and more accurate than other methods mentioned in this paper. The method will have broad application prospects through continuous efforts and improvements.