Comparison of permissible source region and multispectral data using efficient bioluminescence tomography method.

Chenghu Qin,Shouping Zhu,Jinchao Feng,Jianghong Zhong,Xibo Ma,Ping Wu,Jie Tian
DOI: https://doi.org/10.1002/jbio.201100049
IF: 3.3899
2011-01-01
Journal of Biophotonics
Abstract:As a novel molecular imaging technology, bioluminescence tomography (BLT) has become an important tool for biomedical research in recent years, which can perform a quantitative reconstruction of an internal light source distribution with the scattered and transmitted bioluminescent signals measured on the external surface of a small animal. However, BLT is severely ill-posed because of complex photon propagation in the biological tissue and limited boundary measured data with noise. Therefore, sufficient a priori knowledge should be fused for the uniqueness and stability of BLT solution. Permissible source region strategy and spectrally resolved measurements are two kinds of a priori knowledge commonly used in BLT reconstruction. This paper compares their performance with simulation and in vivo heterogeneous mouse experiments. In order to improve the efficiency of large-scale source restoration, this paper introduces an efficient iterative shrinkage thresholding method that not only has faster convergence rate but also has better reconstruction accuracy than the modified Newton-type optimization approach. Finally, a discussion of these two kinds of a priori knowledge is given based on the comparison results.
What problem does this paper attempt to address?