A simulation and estimation framework for intracellular dynamics and trafficking in video-microscopy and fluorescence imagery.

Image sequence analysis in video-microscopy has now gained importance since molecular biology is presently having a profound impact on the way research is being conducted in medicine. However, image processing techniques that are currently used for modeling intracellular dynamics, are still relatively crude and yield imprecise results. Indeed, complex interactions between a large number of small moving particles in a complex scene cannot be easily modeled, limiting the performance of object detection and tracking algorithms. This motivates our present research effort which is to develop a general estimation/simulation framework able to produce image sequences showing small moving spots in interaction, with variable velocities, and corresponding to intracellular dynamics and trafficking in biology. It is now well established that spot/object trajectories can play a role in the analysis of living cell dynamics and simulating realistic image sequences is then of major importance. We demonstrate the potential of the proposed simulation/estimation framework in experiments, and show that this approach can also be used to evaluate the performance of object detection/tracking algorithms in video-microscopy and fluorescence imagery.