Differentiable adaptive optics

Another publication on differentiable adaptive optics.
Paper in Optics Express
Github

Figure: Conceptual scheme of differerntiable adaptive optics.

Inspired by differentiable rendering and physics - here we discover the sample aberration by optimizing a forward model of image formation in the microscope. We match the model to the actual (aberrated) measurements from the actual microscope using gradient-based optimization within a machine learning framework (Tensorflow).

By changing the target function, this approach can be adapted to other tasks, such as generation of holorgrams. For further details check the publication, it is open access, as well as the project repository.

Figure: Examples of computer generated holograms with differentiable model optimzation.