Bioluminescence Tomography by an Iterative Reweighted ${\bm {l_{2}}}$-Norm Optimization

Bioluminescence tomography is a promising tool in preclinical research, enabling noninvasive real-time in vivo imaging as well as quantitative analysis in small animal studies. Due to the difficulty of reconstruction, continuous efforts are still made to find more practical and efficient approaches. In this paper, we present an iterative reweighted l2-norm optimization incorporating anatomical structures in order to enhance the performance of bioluminescence tomography. The structure priors have been utilized to generate a heterogeneous mouse model by extracting the internal organs and tissues, which can assist in establishing a more precise photon diffusion model, as well as reflecting a more specific position of the reconstruction results inside the mouse. To evaluate the performance of the iterative reweighted approach, several numerical simulation studies including comparative analyses and multisource cases have been conducted to reconstruct the same datasets. The results suggest that the proposed method is able to ensure the accuracy, robustness, and efficiency of bioluminescence tomography. Finally, an in vivo experiment was performed to further validate its feasibility in a practical application.