Simulated biological cells for receptor counting in fluorescence imaging

Digital image processing and epi-fluorescence microscopy provide one of the main and basic tools for living biological cells analysis and studying. Developing, testing and comparing those image processing methods properly require the use of a controlled environment. A verified and trustworthy database of images and meta-information is needed to control the validity of the processing results. Manually generating that golden database is a long process involving specialists being able to apprehend and extract useful data out of fluorescent images. More and more, we need to automate this process. Having enough cases in the database to challenge the processing methods and gain trust in them cannot be realistically be achieved manually. This paper presents a framework implementing a novel approach to generate synthetic fluorescent images of fluorescently-stained cell populations by simulating the imaging process of fluorescent molecules. Ultimately, the proposed simulator allows us to generate images and golden data to populate the database, thus providing tools for the development, evaluation and testing of processing algorithms meant to be used in automated systems.