Research On Image Reconstruction For The Double Modality System Of DOT And Micro-CT

Diffuse optical tomography(DOT) is a kind of high depth biomedical imagingtechnology which make use of scattered photons to detect the tissue with merit of low costand no radiation damage, hence it is outstanding in the disease diagnosis, pharmaceuticalresearch, monitoring pathology progression and therapies and other biomedical researchfields. However, the inverse problem of DOT is ill-posed and non-linear due to the diffusivenature of the photons inside the biological tissue, which leads to a poor spatial resolution.Multi-modality imaging method can improve its spatial resolution.In this paper, we present a triangle mesh generation method from CT images, whichuses an adaptive boundary sampling algorithm to convert the image of CT into a planarstraight line graph and mesh it with Delaunay method firstly. Then we refine the initial meshby edge subdivision method and relax and smooth the mesh to improve its quality at last.Nodes and elements are labeled during the mesh generation and revised by a revisingalgorithm.We present an image reconstruction method for DOT and micro-CT double modalityimaging system. The method is based on diffuse approximation, and takes use of finiteelement method(FEM) to solve steady-state diffuse equation, Newton type method toreconstruct the optical coefficients and Laplace regularization to incorporate the structuralpriori information. The simulations show that the incorporation of structural prioriinformation improves the quality of DOT greatly.