Computational aspects and perspectives of novel modalities for visible and nonvisible imaging

Abstract

Computational imaging overcomes the limitations of conventional imaging in both visible and nonvisible regimes. It utilizes highly-efficient computer algorithms — from compressed sensing to machine learning — integrated with imaging hardware to produce high-quality images using relatively low-cost setups. Single-pixel imaging, for example, uses only one detector to image a target scene by correlating the detector signal with modulated light patterns. Computational microscopy setups, specifically ptychography-based ones, are now capable of imaging biological samples that were previously optically irresolvable without the need for high numerical aperture lenses. This talk explores the mechanisms of these unconventional imaging systems that are highly adaptable and performs accurate imaging at any wavelength.