Research On CT Image Segmentation Of Aortic Aneurysm Based On Deep Learning Algorithm

Deep Learning is currently widely used in many areas of scientific research,and it has also been used in industrial production,and the effect is excellent.In recent years,with the developing of deep learning algorithms,it has been used in medical image analysis.The advent of deep learning algorithms makes it easier for computers to process large amounts of image data,processing data faster than before.The number of deep learning model is more and more.According to the realization of different functions,deep learning models are divided into many kinds.For example,deep convolution neural networks(CNNs)are applied to image classification,the full convolution Networks(FCN)are applied to segment images,and different network models have different structural parameters.It can be said that the applications of deep learning algorithms are very extensive.Medical image segmentation is that divided medical image into useful areas;this area does not intersect each other,quantitative or qualitative analysis of the disease.Medical Image segmentation in the image processing process has a critical position;it is to achieve the image of the target area of the extraction and quantitative representation basis of the organization.In this paper,we will focus on using deep learning algorithm segment aortic aneurysm CT image and OCT choroidal,which is introduced as follows:(1)To solve the problem that an aortic aneurysm can be separated from CT images,we design a TJ-1 network model based on the deep learning algorithm,use our existing data sets,combine with our built-in deep learning algorithm platform to achieve training the TJ-1 network model.Finally,through the experimental verification,we prove that the experimental process we design is very efficient.(2)Based on the TJ-1 model,we create the TJ-2 model;we use the TJ-2 model to learn the OCT image edge weight and then combine with the graph search algorithm to complete the OCT choroid segmentation experiment.Through the experiment,we find that we mention the method,have good application prospects,it is worth to continue to study.We designed the experimental program;the segmentation effect is real,time efficiency is high.