Based On L1 Norm Sparse Fluorescent Molecule Tomography Reconstruction Algorithm

Fluorescence molecular tomography (FMT) is a new modality of optical imaging that is expected to play an important role in disease early diagnosis and drugs development. However, to fulfill3D reconstruct in FMT, reconstruction algorithms have to figure out a highly ill-posed problem. In this contribution, we propose efficient optimization algorithms to solve the large-scale reconstruction problem for FMT, in which the sparsity of fluorescent targets is taken as useful a priori information in designing the reconstruction algorithm.First, we utilize spare regularization algorithm based on adaptive finite element method for reconstructing the fluorescent target. Because the fluorescent target is sparse for muscle region in the coarse mesh, we use AMP to reconstruct the target and the goal of us is achieving the range position of it. We regard the result of reconstruction on the first mesh as permissible region, and it will act as the boundary information for reconstruction on the refined mesh. In the refined mesh, fluorescent target is not sparse for permissible region, therefore, we adopt PALM to reconstruct the fluorescent target and obtain the final result. Simulation experiments and phantom experiments show that our proposed method is accurate and robust for FMT.Second, a fast sparse approximation scheme combined with a stage wise learning strategy called Stagewise fast LASSO(Swf_LASSO) enable the algorithm to deal with the ill-posed inverse problem at reduced computational costs which is inspired by the optimization algorithms in the field of compressed sensing. We validate the proposed fast iterative method with numerical simulation on a digital mouse model. The reconstruction results are evaluated with position error, time cost and the fluorescent yield. Compared with the counterparts, our fast iterative method performs better in both imaging quality and computational efficiency. The experimental results demonstrate that our proposed method yield accurate and stable reconstruction from insufficient and contaminated measurements, and it is robust to noise and finite element discretization.