Ptychographic inversion algorithms for intensity data in various experimental configurations.

The project is about new applications of a computational imaging technique called ptychography.
Ptychography was originally a solution to the phase problem in diffraction patterns. Similar to holography, it can measure the phase of scattered waves as well as just their intensity.
However, the method is now seen to have much wider applications in situations where only images can be measured. Other developments mean that it can also cope with what are called 'multiple modes' in an imaging experiment (two unrelated image phenomena occurring in a single data set) such an oscillating 'mixed state'.
The method uses variants on the same core inversion algorithm used in 'conventional' ptychography. The specific aims of the project will be to understand the limits of the performance of these algorithms in 'intensity-type' ptychographic inversion.
Applications of this generic technique could include improving the resolution performance of astronomical telescopes, a new way of tackling the non-isoplanic blind deconvolution problem, and using its modal properties for message encryption. If calculations are successful, the algorithms will be tested on model data collected on the optical bench.
To date, initial worked has centred on a fundamental asymmetry in ptychography. In 'real-space' ptychography, the intensity of diffraction patterns is measured, in 'Fourier ptychography' the intensity of images is measured. Via the principal of reciprocity, both types of data should contain identical information, whereas in fact the quality of the phase and modulus reconstructions in these different modes is very uneven (good quality phase, poor modulus in real-space ptychography, and vice versa for Fourier ptychography). We have determined the reason for this and are now trying to rectify the relevant reconstruction algorithms to overcome this fundamental limitation in the technique.