Compressed computational imaging based on optical differentiation

In computational imaging, getting better imaging quality with shorter time usage is always a challenging problem. The powerful compressed sensing functions as a backend algorithm, which leaves room for us to develop a methodology of compression in imaging systems. Optical differentiation was widely utilized in direct imaging to highlight the features of an image. We apply optical differentiation to compress information in the correlated imaging system. The experimental results indicate a significant improvement in the signal-to-noise ratio and imaging speed. In addition, this scheme enables phase imaging from the second-order correlation. Our work can spark potential applications in biological microscopic and scattering media imaging.