Molecular imaging (MI) is a newly emerging and rapidly developing biomedical imaging field in which the modern technologies and instruments are being married to in vivo represent biological and medical processes directly, sensitively, non-invasively and specifically as well as diagnosing and managing diseases better at cellular and molecular levels. It provides an effective information acquisition and assay methodology for observing function of specific genes, origin and development of living subjects and diseases, treatment effect and dynamic change of drug. Among molecular imaging modalities, optical imaging has become an effective tool in life science field for its excellent performance, non-radiativity and high cost-effectiveness compared with traditional imaging techniques like X-ray computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET) etc. As an important optical molecular imaging technique, bioluminescence tomography (BLT) is deeply studied in this thesis. In general, the goal of BLT is to perform a quantitative reconstruction of an internal bioluminescent source distribution with the scattered and transmitted NIR light signals measured on the external surface of a living small animal or a physical phantom. However, BLT is severely ill-posed for high scattering properties of the biological tissues and the limited and noisy boundary detection data. Therefore, the unique and quantitative reconstruction of bioluminescent source and the development of a fast and robust tomographic algorithm are topics for further investigation. In order to lay a solid foundation for the application of BLT in cancer diagnosis, drug discovery and development fields, BLT is researched from the following aspects: photon propagation, source reconstruction and prototype system. The main contributions of this thesis include following issues: 1. We propose Galerkin-based meshless methods to solve photon transport problem in the heterogeneous biological tissue. In comparison with conventional FEM analysis or MC approach, these algorithms do not require any mesh to discretize the problem, in which the photon flux density is approximated entirely in terms of a group of scatter nodes, and the discrete equations are constructed without consideration of element connectivity information and interrelationship among the nodes. Furthermore, the adaptive technology can be performed easier than FEM. the feasibility of the proposed methods has...
修改评论