Research On Improved Region Growth Method For CT Image Segmentation And 3D Reconstruction

At present,the research on image processing of medical equipment has become one of the frontier topics of science and technology.In the application of image processing technology in medicine,the requirements for the accuracy and speed of image processing results will be more stringent than those in other fields,because the necessary prerequisite for most medical treatment is to accurately obtain the image of the lesion area to analyze the condition and facilitate the formulation of subsequent medical plans.The accuracy of the image of the lesion area has a great impact on the success rate of surgery.Especially in the case of liver cancer,lung cancer and other diseases with the highest mortality of cancer patients all year round,the obtained CT image will be the main tool to observe the diffusion degree of cancer cells in focus organs,and the final imaging error will affect the life safety of patients.In this paper,taking the reconstruction of the three-dimensional image of the liver tumor as an example,the two-dimensional CT image of the liver tumor with noise is selected from the CT database.Firstly,the format is transformed into an image format that can be recognized by MATLAB.Aiming at the factors of uneven gray value distribution of the noise image,the median filter algorithm is selected to preprocess the image,so as to optimize the uneven gray value distribution of the CT image caused by noise.Aiming at the gap problem of image spacing after imaging,the linear weighted average interpolation method is used to interpolate the image and optimize the three-dimensional reconstruction results.Based on the improved region growth algorithm proposed in this paper and the advantages of the two segmentation methods,the adaptive segmentation of two-dimensional CT image is realized,and the segmentation efficiency is improved.A variety of evaluation functions are used to compare the segmentation results with those of other segmentation methods to demonstrate the feasibility of this algorithm.Finally,through the optimized MC algorithm,the final three-dimensional reconstruction of the image is carried out to obtain the final three-dimensional liver tumor image.It mainly has the following innovations:(1)Analyze the noise source of CT image at the present stage.According to the demand for uniform gray value of CT image in this paper,properly filter the CT image to make the gray value of the image as uniform as possible while retaining the image edge.At the same time,the linear weighted average interpolation method is used to interpolate the CT slices to optimize the problem of large spacing of each slice in three-dimensional reconstruction.(2)A rigid visceral tumor segmentation method based on the combination of regional growth algorithm and clustering algorithm is proposed to realize adaptive liver tumor segmentation.The steps of manually selecting seed nodes,manually setting growth criteria and later secondary segmentation processing in the original algorithm are omitted,which speeds up the segmentation efficiency.At the same time,combined with filter preprocessing,the overall segmentation accuracy is significantly improved compared with the original algorithm.(3)Optimize the traditional MC algorithm,solve the error problem caused by the existence of anisotropy and redundant polygons of MC algorithm,cooperate with the previous interpolation processing,and finally get the three-dimensional reconstruction results.Through the evaluation function,the reconstruction time is shorter than the original algorithm,and the accuracy of three-dimensional reconstruction is more accurate than the original algorithm.