Analysis And Inversion Of Abnormal Heat Source Information In Vivo Based On Biothermal Conduction

Chronic non-communicable diseases have always been a disease with high global mortality rates.One contributing factor is that lesions go undetected early in the latent period and crucial treatment/medical intervention time is often missed thereby increasing risks to patients and reducing their survival chance.Current imaging diagnostic equipments have their limitations.To cite an example,they can only diagnose when the organizational structure changes.A medical infrared thermograph comes with a more imperative role to reduce such limitation.It can detect abnormalities when functional changes occur within tissues and realize early detection of lesions thereby promoting early identification and intervention.This is a major advantage for medical infrared thermograph in the field of diagnosis.However,a medical infrared thermograph only achieves qualitative diagnosis and cannot accurately diagnose the lesion.There is a greater demand for a medical infrared thermograph to accurately locate lesions.Based on the theory of biothermal conduction,this paper explores the analysis of abnormal heat sources and the inversion of heat source parameters.As an example,based on the forearm tissue,a uniform tissue model embedded in the point heat source or the ball heat source and an anatomical physiological model embedded in the ball heat source can established.Different heat source parameters are analyzed by ANSYS and simple experiment combined with body surface temperature distribution and radial temperature distribution across the heat source.Finally,the inverse problem in biothermal conduction is solved by analytical method and BP neural network respectively that is,the heat source information parameters are inverted by the surface temperature distribution.The main results of this paper are as follows:(1)Based on the forearm tissue this paper analyzes the biothermal conduction process,establishes a uniform tissue model,an anatomical model and its boundary conditions and briefly describes the steps of the ANSYS finite element method for heat conduction analysis.(2)Through the ANSYS and experimental analysis of different heat source parameters the heat source depth has the greatest influence on the surface temperature distribution and the heat source temperature has the least influence.In the analysis of the radial temperature distribution the heat source depth does not affect the heat conduction process in the uniform structure.In the anatomical physiological model due to the interference of other information the change of heat source depth will affect the heat transfer process.(3)This paper uses the analytical method to invert the heat source information in the uniform tissue model.The relative error of the analytical method for depth inversion of point heat source is about 10% and the relative error of ball heat source information inversion is less than 14%.The inversion effect of the heat source temperature is the best but in the inversion process when the diffusion radius increases the accuracy of the heat source parameter estimation will be reduced.(4)In this paper a BP neural network is designed based on the characteristics of biothermal conduction to realize the inversion of heat source information by body surface temperature distribution.The inversion error of the heat source parameters under the uniform tissue model is within 9%,and the error under the anatomical physiological structure model is within 18%.It is feasible to use BP neural network to solve heat source information.Under the same model the inversion errors of different heat source parameters are not much different and more objective inversion and analysis can be achieved for different heat source parameters.