Optimizing near-infrared polariscopic imaging for the living human eye

Abstract

Hardware architectures and image interpretation can be simplified by partial polarimetry. Mueller matrix (MM) polarimetry allows the investigation of partial polarimeter designs for a given scientific task. In this work, we use MM measurements to solve for a fixed polarization illumination and analyzer state that maximize polariscopic image contrast of the human eye. The eye MM image acquisition takes place over 15 seconds which motivates the development of a partial polarimeter that has snapshot operation. Within the eye, the birefringent cornea produces spatially-varying patterns of retardance exceeding half of a wave with a fast-axis varying from linear, to circular, and elliptical states in between. Our closed-form polariscopic pairs are a general solution that maximizes contrast between two non-depolarizing pure retarder MMs. For these MMs, there is a family of polariscopic pairs that maximize contrast. This range of solutions creates an opportunity to use the distance from optimal as a criteria to adjust polarimetric hardware architecture. We demonstrate our optimization approach by performing both Mueller and polariscopic imaging of an in vivo human eye at 947 nm using a dual-rotating-retarder polarimeter. Polariscopic images are simulated from Mueller measurements of 19 other human subjects to test the robustness of this optimal solution.