Abstract:Imaging through scattering media has become an important aspect in optical imaging. To achieve this, different approaches have been presented to recover the phase of an imaging object from its scattered speckle behind a scattering medium. However, all these algorithms require initial parameters which are obtained based on prior knowledge of the imaging object. This is impractical, as prior knowledge of the imaging object is not always available in real imaging applications. In this paper, we propose using a Particle Swarm Optimization algorithm to find these optimal parameters without requiring prior knowledge of the imaging object. These optimal values are then adopted in the Phase Retrieval algorithm based on object-modulated speckles for object image reconstruction behind a scattering medium in a single-shot imaging system. The obtention of the values for the standard deviation for Gaussian filter and the threshold control allows Phase Retrieval to converge faster and avoids the need for running the algorithm more than once in the search for the best reconstruction.

Tuning parameters of phase retrieval algorithm for single-shot imaging based on object-modulated speckles by particle swarm optimization