Exploiting data redundancy in computational optical imaging.

We present an algorithm which exploits data redundancy to make computational, coherent, optical imaging more computationally efficient. This algorithm specifically addresses the computation of how light scattered by a sample is collected and coherently detected. It is of greatest benefit in the simulation of broadband optical systems employing coherent detection, such as optical coherence tomography. Although also amenable to time-harmonic data, the algorithm is designed to be embedded within time-domain electromagnetic scattering simulators such as the psuedo-spectral and finite-difference time domain methods. We derive the algorithm in detail as well as criteria which ensure accurate execution of the algorithm. We present simulations that verify the developed algorithm and demonstrate its utility. We expect this algorithm to be important to future developments in computational imaging.