Trust Region Method for Solving the Bioluminescence Tomography Inverse Problem

Abstract

Bioluminescence Tomography (BLT) reconstruction is a highly ill-posed inverse problem. In this research, we formulate it as a quadratically constrained least squares problem and present a trust region method for solving this problem. Simulations based on a heterogeneous phantom demonstrate that the trust region method has the capability to effectively handle the ill-posedness of BLT. The experiments results show that the proposed method outperforms Tikhonov regularization in terms of source location and reconstructed source energy.