Understanding the Gerchberg-Saxton Algorithm through Convex Optimization

[j-lee00]

Please help me understand the Gerchberg–Saxton algorithm

what I comprehended so far

1) A source is shone on an object and produces a diffraction pattern in the diffraction plane. We do not know the dimensions of the object.

2) We are unable to calculate the phase of the of the wave, due to the "phase problem"

Now we want to computationally find the phase of the wave that is produced in the diffraction plane


I do not understand how the Gerchberg–Saxton algorithm works. i.e. how we improve to get a closer phase through the iterative method

Thanks

 

[Wrichik Basu]

If you have checked Wikipedia and your doubts are still not clear, trying reading this paper available freely at ResearchGate:

https://www.researchgate.net/publication/11281090_Phase_retrieval_Gerchberg-Saxton_algorithm_and_Fienup_variants_A_view_from_convex_optimization