In vivo and in situ cellular image processing and characterization:: Challenges and solutions

We study the feasibility of 3D virtual histology through in vivo and in situ cellular imaging. A prototype system has been developed based on photodynamic fluorescence signals, confocal endomicroscopy, and FPGA image processing and characterization computing. Experiments in its clinical applications have been conducted, mainly for diagnosis of early-stage mucous cancer. With the fine-grained parallel imaging programs mapped on the FPGA, a stream of focused optical sections of microstructures in the subsurface layers up to 300 mu m in depth, can be processed online and the extracted features can be visualized seamlessly with the endomicroscopy settings.