Dynamic Inverse Problems in Time-Lapse Microscopy

Recent advances in microscopy technology have enabled high-resolution imaging of biological processes at the molecular level. However, the dynamic nature of these processes and low signal-to-noise acquisition pose challenges such as tracking of objects with complex dynamic behavior and ill-posed image reconstruction algorithms. The project addresses these challenges by investigating dynamic inverse problems in time-lapse microscopy and proposing an algorithmically efficient and stable solution with temporal structural models that constrain object motion and topology changes. Specific goals include robust reconstruction of high-resolution data, multi-object detection and tracking, fitting and learning object’s structural models, and testing of the developed algorithms on biomedical problems. The project has the potential to significantly advance automated processing in microscopy imaging and our understanding of complex biological processes. The proposed solution will be applied to data from super-resolution light microscopy and environmental scanning electron microscopy.