Simplified retinal 3D projection rendering method and system

Abstract

A simplified rendering method and system for retinal 3D projection using view and depth information is proposed and demonstrated. Instead of vertex calculations, image-based techniques, including sub-image shifting, image fusion, and hole filling, combined with the depth information, are used to render the multi-view images in a display space with specific discrete depth coordinates. A set of time–division multiplexing retinal 3D projection systems with dense viewpoints is built. A near-eye display of a 3D scene with complex occlusion relationships is realized using the rendering method and system. The eye box of the retinal projection system is enlarged, and the accommodation response of the eyes is evoked at the same time, which improves the visual experience. Rendering tests are carried out using simple and complex models, which proves the effectiveness of this method. Comparative experiments prove that the proposed retinal projection method can obtain high-performance 3D images comparable to the super multi-view display method while simplifying the rendering process. Additionally, the depth of field of the experimental system can cover most of the vergence accommodation conflict sensitive range of the human eye.