Three-dimensional quantitative optical microscopy imaging applied to large scale digital cell analysis system

The Large Scale Digital Cell Analysis System (LSDCAS) is a cell imaging and analysis system applying automated fluorescence microscopy and phase contrast microscopy. Non-perturbing cell culture conditions can be maintained in LSDCAS. This distinct feature makes LSDCAS an extensible live cell imaging platform. This thesis describes techniques developed to provide 3-D quantitative imaging capabilities for LSDCAS. Deconvolution techniques have been widely used for restoring 3-D quantitative information of an unknown specimen using fluorescence microscopy. Three constrained iterative deconvolution algorithms: one based on a Poisson noise model and two others based on a Gaussian noise model, are presented and were evaluated using synthetic images and experimentally obtained microscope images. Based on them, an open-source deconvolution package named as Deconv was implemented and released under the GNU Public License. This package provides a library which includes C++ classes for the simulation of a 3-D point spread function and 3-D deconvolution of optical sectioning images acquired from a fluorescence microscope. For phase contrast imaging, deconvolution is not applicable due to the inherent non-linear relationship between the specimen properties and image intensities. A computational model applied to phase contrast microscopy imaging using ray tracing techniques was built. This model was then evaluated and validated using a synthetic specimen and a latex microsphere. This model will be used to develop techniques to reconstruct unknown specimens from optical sectioning images acquired from a phase contrast microscope.