Study On The Image Reconstruction Algorithms For Diffuse Correlation Tomography Of Blood Flow

Many diseases are related to abnormal blood flow values in tissues.Near-infrared diffuse correlation tomography(DCT)is an emerging technology for non-invasive imaging of biological tissue blood flow,in which the image reconstruction algorithm is a key factor in the effect of blood flow imaging.This article has carried out a more in-depth and systematic research on the DCT blood flow reconstruction algorithm,focusing on investigating and comparing two types of reconstruction algorithms,namely the finite element method(FEM)representing the partial differential equation imaging framework and the integral equation imaging framework Nth-order linear(NL)algorithm,the main research content includes:First,DCT reconstruction algorithm based on finite element(FEM)method.Using the open source software package Nirfast based on diffuse optical tomography(DOT),combined with the differential equation expression of DCT,the finite element modeling and blood flow imaging calculation are realized.Second,position design of optical probe(S-D).In order to make full use of the Monte Carlo(MC)photon information(eg.photon path,weight)simulated to improve the quality of blood flow reconstruction,this paper designed a new optical probe S-D array.Third,Establish computer simulation models of the above two DCT reconstruction algorithms;and carry out phantom experiments to quantitatively evaluate the reconstruction algorithms.In order to evaluate the blood flow imaging quality of the FEM and NL models,this paper defines three parameters,Mean Error,Uniformity(CV)and Contrast(?),to evaluate the reconstruction results.Both computer simulation and phantom experiment results show that under the same S-D array and spatial voxel conditions,the NL reconstruction algorithm issuperior to the FEM method in terms of contrast and uniformity.It shows that DCT blood flow imaging is greatly affected by the reconstruction algorithm,so it has great potential in future algorithm design and optimization.