Foveated Retinal Optimization for See-Through Near-Eye Multi-Layer Displays

In order to implement 3-D displays with focus cues, several technologies, including multi-layer displays, have been introduced and studied. In multi-layer displays, a volumetric 3-D scene is represented by 2-D layer images via optimization process. Although this methodology has been thoroughly explored and discussed in optical aspect, the optimization method has not been fully analyzed. In this paper, we deal with pupil movement that may prevent efficient synthesis of layer images. We propose a novel optimization method called foveated retinal optimization, which considers the foveated visual acuity of human. Exploiting the characteristic of human vision, our method has tolerance for pupil movement without gaze tracking while maintaining image definition and accurate focus cues. We demonstrate and verify our method in terms of contrast, visual metric, and experimental results. In experiment, we implement a see-through near-eye display that consists of two display modules, a light guide, and a holographic lens. The holographic lens enables us to design a more compact prototype as performing the roles of an image combiner and floating lens, simultaneously. Our system achieves $38^\circ \times 19^\circ $ field of view, continuous focus cues, low aberration, small form factor, and clear see-through property.