Modeling And Applications Of Free-space Light Propagation In Optical Molecular Imaging

Noncontact optical tomography (NOT) has attracted increasing attentions in recent yearsbecause of its significant advantages. As a promising optical molecular imaging technology,NOT is dominant in sensitivity, operation, cost-effectiveness and safety. Furthermore, NOTremoves the use of the optical fibers, imaging holder and matching fluid and avoids thecontact between the signal detection module and object. As a result, it simplifies the systemframework and improves the convenience for the system development and experimentalmeasurement. Because of the use of the high sensitivity CCD camera and camera lens, NOTcan capture optical signals with high resolution so that it improves the sensitivity, resolution,and image quality. However, the lack of free space light transport theory brings a greatdifficulty and challenge for the development of NOT technique. In this dissertation, weinvestigated the modeling and applications of free space light transport theory. The maincontribution of this dissertation can be summarized as follows:1. Based on the simplification theory of camera lens and Monte Carlo method (MCM),we studied the MCM based free space light transport model and the MCM based simulationframework for the forward problem of NOT. Considering the poor simulation efficiency ofMCM, we implemented parallel accelerated model for the forward simulation frameworkbased on small parallel system. Comparisons validated the accuracy and effectiveness of theMCM based free space light transport model and multi-CPUs based simulations demonstratedthe parallel performance of the parallel simulation framework.2. We studied hybrid radiosity-radiance theorem based free space light transport model.Based on the simplification theory of camera lens, we constructed a simplified model fornoncontact optical imaging system. Then, the spatial location relationship between the bodysurface and CCD camera was established by introducing a virtual detection plane and avisibility factor. Finally, the mapping process of light from body surface to the CCD camerawas obtained by using the Lambertian source theory of diffuse light and the radiance theoremof image. Compared with experimental measurements demonstrated the accuracy andeffectiveness of the developed model.3. When one simulates the detection results on left or right perspective using thedeveloped hybrid radiosity-radiance theorem based free space light transport model, there is alittle difference compared with experimental measurements. Analyzing the potential reasonsof this difference and considering the effects of the camera lens diaphragm on detectionresults, we proposed an improved free space light transport model based on the analysis ofcamera lens diaphragm. Compared with experimental measurements, we validated the accuracy and effectiveness of the improved model. And comparisons with our previous modelverified its improved performance.4. Motivated by the principle of the optical path reversibility and the ideology of thereciprocity theorem between the sources and detectors, we investigated the inverse problem ofthe free space light transport model and proposed a novel and effective method thatimplements three-dimensional reconstruction of light flux distribution on arbitrary surfacefrom two-dimensional multi-photographic images. The accuracy and effectiveness of theproposed method were validated with comparisons involving both regular and irregularsurfaces. A living mouse based experiment demonstrated the potential of the proposed methodin its applications in optical tomography.5. We investigated the applications of the inverse problem of free space light transportmodel in bioluminescence tomography (BLT). On account of the need of the anatomicalstructure when one performs source reconstruction in BLT technique, we presented adual-modality bioluminescence tomography/micro computed tomography (BLT/uCT)imaging framework and demonstrated its feasibility and potential with experimentalmeasurements. Moreover, because some abstruse problems appears if a structural imagingtechnique was used to provide the anatomical structure information, we proposed anall-optical BLT imaging framework to avoid the use of the structural imaging techniques.Compared with experimental measurements, we validated the feasibility and effectiveness ofthe all-optical framework.6. We investigated the application of the free space light transport theory in endoscopicoptical tomography (EOT) technology, and proposed a novel endoscopic algorithm for EOT.EOT technique can provide accurate images of large animals or human body by taking thedetector closer to the region of interest through an organ cavity with a minimally invasivemanner. Firstly, by analyzing endoscopic geometry of EOT, we constructed an accurate hybridlight transport model. Then, an effective source reconstruction algorithm was developed byusing the finite element method and the sparse regularization technique. A series of bigheterogeneous phantom based simulations validated the accuracy and effectiveness of theproposed algorithm, and also demonstrated the potential of the NOT technique in imaging ofthe internal organ inside a large animal or human body.