Image Segmentation Algorithm Of Liver Anatomy

In the treatment of liver diseases,CT images can provide doctors with a strong basis for diagnosis.At the same time,it can also play a key role in determining the treatment plan and surgery.However,in the process of searching for the liver contour,we have to face many problems such as unclear liver edges,close gray scale of liver and other organs,and tumors inside the liver.This has brought great difficulties for us to find liver contours.Therefore,this thesis proposes an idea of finding the liver contour in a CT image.The idea is that each pixel of the CT image corresponds to a binary parameter,when_x=1 represents the interior of the liver and_x=0 represents the background area,the list of binary parameters is used as the independent variable of the loss function.We will expand on the given multiple initial areas by reducing the value of the loss function,use the Canny method to find the edges of all parts of the liver CT image,and find all edges in the expanded area finally.The edge found by this method is that we are looking for the edge of the liver.The expansion rule provided at the beginning of this thesis can only process part of CT images.We changed the original expansion stop rule by looking for the change rule of the minimum value of the five loss functions in the expansion direction,and solved the question that the regional expansion terminated prematurely.This method also has good results for the complicated internal liver CT images.The combination of these two methods solves the shortcomings that the Canny method can only obtain all the edges,but it can’t obtain the edges of the required parts.It also avoids the unacceptable amount of calculation caused by only using the loss function to calculate the pixel by pixel.