Final Report Summary - ICON (Integrated Real-time Feedback Control and post-processing for image Restoration) The challenging application of doing data driven modelling and control for real-time image resolution enhancement in high performance imaging instrument has been a research driver of a three main new research developments. The first is the development of a new methodology to model such large network connected dynamical systems using data or measurement acquired from the imaging instrument. The research challenge to resolve the curse of dimensionality of these large models was the development of new model structures that only require sparse model parametrization on the one hand and the development of methods to find these parameters based on measured data acquired from these networked systems on the other hand. Such novel identification methods were developed in the scope of iCON. These new methods have the appealing property that for large network connected systems they are either able to find estimates of the model parameters with a computational complexity that is linear in the size of the network or they are able to identify local models operating in a global network using only local measurements of that local system and those within a small neighbourhood. And this while the local system is communicating with neighbouring systems under the condition that this inter system communication is unknown. Such problems easily in network connected systems when only (non-measurable) state information is exchanged between local systems, such as for example in discretizing distributed parameter systems governed by partial differential equations. The second is the development of new computational methods for the reconstruction of wavefront aberrations that reduce the image quality in high resolution imaging instruments. These new computation methods allows to handle important practical constraints, such as sparsity in the parametrized wavefront, both high and low Numerical Aperture problems and both when imaging point sources and extended objects. Special attention is again given to the efficiency of the computational procedures to be able to handle large scale problems. These are fundamental contributions to the research area known as phase retrieval. The third is the development of data driven controller design methods that are able using measurements of imaging systems, acquired through standardly available imaging sensors such as Shack-Hartmann, deformable mirrors and science camera’s, to device optimal controllers that attempt to maximize the image quality enhancement when the imaging is conducted on temporal spatial varying aberrations. Such data driven methods enable to optimize the resolution of a high resolution imaging instrument in a defacto automatic manner by an non-optical expert. This outcome will have be a big stimulus for integrating and re-tuning complex Adaptive Optics systems for high resolution instruments using Adaptive Optics, increasing their availability such as in Optical Telescopes, but it will also enhance to dissemination of this (complex) technology to a much wider class of imaging instruments. The latter is demonstrated through the development of a number of demonstrated, one is the proof-of-concept demonstrator Adaptive Imaging Microscopy (AIM) that was developed in the companion POC project. The latter demonstrator was a compact add-on that could be mounted on existing widefield microscopes in order to give them the capabilities of a light sheet microscope. The add-on feature is made possible by the modular design in combination with the feedback philosophy of the iCON project for improving the optical quality.