Adaptive Optics (AO) is required to achieve cellular resolution at high numerical aperture in small animal eyes. Development of AO technology is required to lower the barriers of the technology translation into pre-clinical vision research environments. Aberration correction for retinal imaging has been demonstrated with great results by direct wavefront sensor (WFS) measurement. However, in some cases the performance of WFS-based AO can be limited by several factors including common path errors, wavefront reconstruction errors and an ill-defined reference plane on the retina. Image-based AO can avoid these issues and the cost of algorithmic execution time. We are investigating and evaluating image-based approaches to potentially provide improvements to compactness, accessibility, and performance of AO systems.
Our experiments were performed on a AO-SLO system which relayed the mouse pupil to a continuous deformable mirror (DM) and Shack-Hartmann wavefront sensor on conjugate planes. The system allows for closed-loop AO as well as the ability measure aberrations during the image-based optimization. We characterized our DM with a WFS in order to use open-loop modal control. The image-based optimization searched the Zernike polynomial coordinate system using a sharpness quality metric. Our results demonstrate diffraction limited performance (according to the WFS) with both closed-loop and image-based methods. The number of iterations required for the image-based method is dependent on the aberrations present as well as the number of dimensions being corrected. Image-based optimization after closed-loop correction can provide further improvements. We are applying these results to improve image-based AO for small animal in-vivo applications.