Segmentation Technology Research And Implementation Of Sinusoids In Liver Pathology Image

Liver with complicated structure and high rate of morbidity is the biggest viscus of human's body.Liver disease has become the number one killer disease endangering people's health.Early detection and treatment is the key to cure it.With the constant development of computer technology and iconography theory,it is possible to makes early diagnosis on the liver disease based on the medical iconography.CAD technology is in the ascendant.At the same time,more and more Computer Aided Diagnosis Systems are applied for practical clinical diagnosis.Due to the traditional imaging technology hard to meet people's need,pathology image is born at the right time.Sinusoid is a very important measuring index to liver disease such as liver cancer,liver cirrhosis,and alcoholic liver.Sinusoid segmenting of liver pathology imaging featuring adhesive tissue,complicated structure,similar color,fuzzy boundaries,is the important and hardest sector in the CAD system.How to segment the Sinusoid in the accurate,quick and efficient way has becomes chief difficulty.This paper proposes a sinusoid rough segmentation method,that is,liver pathology image enhancement method,combination improved level set method and chroma segmentation method in the study of liver pathology image based on the literatures at home and abroad.Sinusoid segmentation is composed of image denoising,image enhancement,rough segmentation,precise processing.This paper can be divided into the following parts:firstly modelling a multi-orientation selective Gaussian filter to connect disconnected sinusoid boundaries,secondly enhancing it based on the statistical characteristic of liver pathology imaging so as to make the boundaries clear.Thirdly proposing a combination method of improved level set method and chroma segmentation method in the stage of image rough segmentation;then removing non-sinusoid region like fat droplet,glass cell,partial cell nucleus and ballooning cell of the outcomes of rough segmentation.Accurate segment outcomes of liver pathology image can be achieved in the end.To test and verify the efficiency of this improved method system,the writer conduct an experiment on the Sinusoid segmenting of liver pathology imaging.The outcomes proof that this method perform well on the sinusoid segmenting with smooth appearance.In addition,this good robustness system is suitable for analyzing and diagnosing the liver disease.