Study On Electromagnetic Imaging Of Tumor Applying To Radiomics

The concept of radiomics applied to medical was first proposed by Holland scholars in2012.The radiomics needs a lot of image for building database,image segmentation,feature extraction and classification of image information.If we take radiomics for early cancer detection,It needs a large number of early tumor imaging.In view of the many shortcomings of the traditional promotion methods of cancer detection is not suitable for the detection of the early malignant tumors,so we propose to study the early detection of cancer using microwave imaging technology.This technique relies on electromagnetic parameters significantly different from normal tissue and tumor tissue in early carcinogenesis exists,it has the advantages of low radiation energy of electromagnetic signal,high precision,low cost,suitable for detection in high risk population.Microwave tomography is a kind of active microwave imaging,its essence is a kind of method for solving electromagnetic inverse scattering,the electromagnetic signal reverse organizations have deduced the electromagnetic parameter distribution within the organization.Considering the advantages of modern optimization algorithm is better than traditional optimization algorithms,this paper choose to use genetic algorithm to solve the inverse scattering problem of electromagnetic,electromagnetic signal to fit the organization through peripheral electromagnetic parameters iterative tissues distribution.Firstly,we introduce the implementation principle of genetic algorithm,the advantages and the disadvantages,detailed explain how the genetic algorithm will be introduced to the use of microwave tomography.In the simulation stage,we use the proposed algorithm for the early tumor model with simple small size in first,then a simple early tumor model with a larger size was introduced as a comparison,simple model here refers to the normal tissue and tumor tissue which have a certain dielectric constant and conductivity at a fixed frequency.Afterwards,we use the confocal imaging algorithm as a comparison to Analysis the problems of the proposed method and the advantages so that the original method is improved,so as to realize the rapid and accurate imaging results.Finally,we will introduce some complex tumor models which the tumor permittivity and conductivity is not definite,the electromagnetic parameters in a dynamic range of image restoration.The proposed algorithm has large amount of calculation and low iterative speed problem in the process of tumor imaging,but after improving,the problem has been well improved,to ensure the imaging quality and its computing speed is greatly improved.