PARALLEL RECONSTRUCTION SOFTWARE FOR IN VIVO FLUORESCENCE TOMOGRAPHY

点击翻译按钮获取中文摘要

This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. Primary support for the subproject and the subproject's principal investigator may have been provided by other sources, including other NIH sources. The Total Cost listed for the subproject likely represents the estimated amount of Center infrastructure utilized by the subproject, not direct funding provided by the NCRR grant to the subproject or subproject staff. This Technological Research and Development project is also collaborative, and is updated to the following: Fluorescence diffuse optical tomography (FDOT) is a noninvasive method of imaging labeled molecules in living organisms, which could revolutionize both basic research and preclinical imaging. Currently, the primary limitation is resolution, and it is widely accepted that the resolution is fundamentally limited by light scattering in biological tissue. A key to successful reconstruction or high resolution is an accurate forward problem solver for light propagation in highly heterogeneous medium, such a mouse. It has been shown in our latest studies as well as in the literature that analytical solutions and finite element methods based on the diffusion approximation are inadequate for free-space FDOT for small animals. In response, we are going to use Monte Carlo (MC) method to solve the forward problem. However, the complexity in geometry and the large number of sampling points in free-space FDOT experiment demand computing resources far beyond what a regular desktop computer can offer. In this project, we are going to work with our collaborators at Pittsburgh Supercomputing Center PSC), and apply parallel computing technology to our previously developed MC-based software packages for FDOT. Specifically, the aims of this project are: (1) Integrate existing software implemented in C (MC simulation) and MATLAB (data processing and image reconstruction) programming languages. (2) Modify the integrated software so that it can utilize parallel computing facilities (supercomputers) at PSC, in order to significantly improve computing performance (one the order of 1000 in terms of computing time reduction) of image reconstruction. (3) Improve the algorithms and implementation of the MC software, so that the simulation/computation results will be more accurate for high-resolution applications. (4) Apply the parallel reconstruction software to FDOT data acquired previously at our lab as well as in future phantom and animal studies. (5) The developed software will be freely available to the public to benefit the scientific community.