A gentle introduction to computational optical imaging [Invited]

Computational imaging has become an important research area in the last few decades. Relaxing the traditional requirement that optical hardware directly form the final image of the object on an image sensor has allowed us to better appreciate the image formation process from an information transfer point of view. The physical properties of light that carry this information are amplitude, phase, polarization, spectrum, coherence, photon correlations, etc. Strategies for efficiently measuring one or more of these information-carrying entities, followed by reconstruction algorithms, then decide the ultimate quality of images obtained from an imaging device. With the advancement in optical manufacturing, the availability of computing power, and the development of novel algorithmic ideas for image reconstruction, it is likely that imaging technologies will continue to evolve rapidly, and as a result, the commonly used imaging devices like cameras, microscopes, and telescopes may take a very different physical form in the coming decades. The performance parameters such as resolution, contrast, field of view, depth of focus, and noise sensitivity will also likely be revised in the process to include both hardware and algorithmic aspects. Apart from pushing the fundamental limits, the synergistic working of optical hardware and algorithms has the additional advantage of making imaging systems economical, resulting in significant social impact. This topical review provides an introduction to basic principles and methods of computational imaging in a manner that is accessible to beginning researchers at the early graduate studies level or industry professionals, in order to get them started in this exciting field. The paper further aims to draw attention of experts in optical sciences, precision engineering, mathematics, computer algorithms, and users in multiple application areas by highlighting the interdisciplinary research and development opportunities.