Abstract

Inverse scattering problems, namely reconstructing the structures of objects from their scattered intensity distributions occur in many fields of science and technology[1], such as tomographic imaging, seismology, single shot X-ray scattering[2] and imaging. Solving the inverse scattering problem where all the phase information is lost, is generally very difficult. The difficulty is alleviated by resorting to some a priori knowledge such as the boundaries within which the object lies (compact support), sparsity or other spatial features. Then it is possible to reconstruct the object using iterative algorithms, such as the well-known Gerchberg-Saxton algorithm. Unfortunately, the algorithms are time consuming and do not always converge to the right solution even with advanced computational resources.

© 2017 IEEE