The Ill-posed Problem Of 3D In-Vivo Fluorescence Optical Tomography Imaging For Small Animal

With the rapid development of targetable fluorescent probe technology and optical imaging technology,3D imaging of in-vivo fluorescence has become the main tool of studying molecular events.This technique has a promising future in scientific researches such as development of new drugs and the evaluation of tumor efficacy.3D imaging of in-vivo fluorescence is a technique that can reconstruct three-dimensional distribution of in-vivo fluorescence based on the information acquired from surface measurement and inner optimal structure parameters of the tissue.However,the limited surface fluorescence information combined with the complex photon transmission in biological tissues leads to the serious ill-posedness of 3D reconstruction.How to improve the accuracy of in-vivo fluorescent sources reconstruction is an important part of this study.The main work of this paper is as follows:?1?We studied the mathematic optimization model of inverse reconstruction problem based on the theory of compressive sensing,and proposed L_p?0<p<1?regularization optimization model that can better highlight the sparseness of in-vivo fluorescence sources.?2?Aiming at the high coherence in columns of coefficient matrix in 3D fluorescence imaging,an orthogonal transform of coefficient matrix based on QR decomposition is proposed,which can effectively improve the imaging accuracy.?3?Aiming at the problem of over-smoothing based on the L₂ regularization optimization model and the non-optimal sparse problem based on the L₁ regularization optimization model,L_1/2regularization-based inverse reconstruction optimization model is proposed and solved iteratively with threshold algorithm.The double light source numerical simulation verifies the effectiveness of the algorithm.?4?The L_p?0<p<1?regularization problem is a non-convex optimization problem.When the number of measurements is small,it will fall into the local optimal problem.In this paper,the L₁ and L_p?0<p<1?alternating iterative algorithm based on simultaneous algebraic reconstruction technique is proposed in order to get the global optimal solution.Nylon imitation experiment verified the effectiveness of the algorithm.?5?Aiming to solve the problem that mice is easy to deform and low reconstruction accuracy,the animal optical tomography imaging system?AOIS?set up by our research group was improved.In-vivo fluorescence experiments of mice demonstrate the effectiveness,stability and robustness of the system.The three-dimensional imaging system of small animals based on compressed sensing theory in this paper can effectively reduce the ill-posedness in inverse reconstruction and improve the imaging quality.