| High-quality and efficient medical treatment can improve people’s quality of life.As a malignant tumor,liver tumors seriously threaten people’s lives and health.Therefore,radiofrequency ablation(RFA)treatment of liver tumors has gradually received widespread attention,and the level of RFA treatment in various regions also needs to be improved urgently.In order to accurately realize the RFA treatment of liver tumors,it is necessary to use CT images to design a needle plan for RFA treatment of liver tumors before surgery.In RFA design,accurate segmentation of liver tumors in CT images is required as an important basis.At the same time,the three-dimensional reconstruction of liver tumors will bring greater convenience to the design of RFA treatment,so it is of great significance to study the three-dimensional reconstruction of CT.This article combines medical image processing and deep learning related methods.First,the liver tumors in the CT image are segmented,secondly,three-dimensional reconstruction of different tissues in the CT image is based on the liver tumor segmentation,and finally based on the established three-dimensional reconstruction model Completed the design of RFA treatment plan for liver cancer.In this paper,the following researches are done to solve the shortcomings of traditional algorithms:(1)In view of the insufficient accuracy of liver tumor segmentation in CT images,this paper proposes a liver tumor segmentation algorithm based on U-net network.The algorithm uses two deep convolutional neural networks that participate in the segmentation task,respectively for the liver segmentation task and liver tumor segmentation task.The whole method is divided into two parts: offline phase and online phase.The network is trained in the offline phase,and in the online phase,the trained neural network is directly used in the CT image segmentation task.Simulation results show that this algorithm can reduce the false segmentation generated in the non-liver region,and effectively improve the segmentation accuracy of liver tumors in CT images.(2)Aiming at the problems of low efficiency of 3D reconstruction algorithm and insufficient practicality of 3D reconstruction model in the process of 3D reconstruction of CT images,this paper proposes a 3D reconstruction algorithm of CT images based on improved Marching Cubes algorithm.The algorithm first increases the volume of voxels by improving the representation of voxels,thereby reducing the number of voxels that need to be traversed during the 3D reconstruction process.At the same time,in the process of calculating the equivalent points,smoothing and distortion coefficients are introduced to control the smoothness and distortion of the 3D reconstruction model.Before drawing the isosurface,the mesh reduction algorithm based on the second measurement error is used to reduce the output triangle mesh,which enhances the real-time interactivity of the 3D reconstruction model.Simulation results show that this algorithm can improve the efficiency of 3D reconstruction and improve the real-time performance of reconstruction.To sum up,this article carried out simulation on the algorithm discussed.Simulation results show that the U-net network-based liver tumor segmentation algorithm proposed in this paper can complete the liver tumor segmentation task and provide a basis for RFA treatment design;the CT image 3D reconstruction algorithm based on the improved Marching Cubes algorithm can be used for subsequent liver tumor RFA treatment Design provides guarantee. |
