There are several ophthalmic devices to image the retina, from fundus cameras capable to image the whole fundus to
scanning ophthalmoscopes with photoreceptor resolution. Unfortunately, these devices are prone to a variety of ocular
conditions like defocus and media opacities, which usually degrade the quality of the image. Here, we demonstrate a
novel approach to image the retina in real-time using a single pixel camera, which has the potential to circumvent those
optical restrictions. The imaging procedure is as follows: a set of spatially coded patterns is projected rapidly onto the
retina using a digital micro mirror device. At the same time, the inner product’s intensity is measured for each pattern
with a photomultiplier module. Subsequently, an image of the retina is reconstructed computationally. Obtained image
resolution is up to 128 x 128 px with a varying real-time video framerate up to 11 fps. Experimental results obtained in
an artificial eye confirm the tolerance against defocus compared to a conventional multi-pixel array based system.
Furthermore, the use of a multiplexed illumination offers a SNR improvement leading to a lower illumination of the eye
and hence an increase in patient’s comfort. In addition, the proposed system could enable imaging in wavelength ranges
where cameras are not available.