Fluorescence microscopy is one of the core technologies for biomedical research and diagnostics. The large diversity of microscopy applications requires sophisticated, individually adapted image processing methods that are able to process huge amounts of data, high background signals as well as low contrast images. The correct identification of objects in micro- or nanometer range is crucial for diagnosing diseases, the prognosing their progress and examining the impact of medical treatments.
In the project NanoDetect we design the foundations of a radically new image processing software framework for biomedical researchers providing high throughput analysis of single molecules in cells. This makes image analysis independent of the researcher‘s personal knowledge and constitution. We develop software modules that combine image analysis functionality with machine learning and pattern recognition algorithms, automatically optimizing the parameters for each data processing step based upon statistical analysis and user feedback. Special consideration is given to the target group of researchers who do not necessarily have computer science knowledge, by adding user friendly interfaces as well as a web service for the NanoDetect framework.