Novel Trace Norm Regularization Method For Fluorescence Molecular Tomography Reconstruction

Fluorescence molecular tomography (FMT) is developing rapidly in the field of molecular imaging. FMT has been used in surgical navigation for tumor resection and has many potential applications at the physiological, metabolic, and molecular levels in tissues. Due to the ill-posed nature of the problem, many regularized methods are generally adopted. In this paper, we propose a region reconstruction method for FMT in which the trace norm regularization. The trace norm penalty was defined as the sum of the singular values of the matrix. The proposed method adopts a priori information which is the structured sparsity of the fluorescent regions for FMT reconstruction. In order to improve the solution efficiency, the accelerated proximal gradient algorithms was used to accelerate the computation. The numerical phantom experiment was conducted to evaluate the performance of the proposed trace norm regularization method. The simulation study shows that the proposed method achieves accurate and is able to reconstruct image effectively.