Resolution enhanced light field near eye display using e‐shifting method with birefringent plate

Abstract

Light field near eye displays (LFNED) can produce lightweight devices and address the accommodation-convergence conflict. However, low spatial resolution creates a poor immersive experience in LFNED. In addition, although many e-shifting devices have been proposed to enhance resolution in projection system, yet those devices are too bulky to be applied in an LFNED to keep it lightweight. In this paper, a compact e-shifting component is proposed to enhance image resolution in an LFNED by using a birefringent plate and twisted nematic switch cell. The proposed e-shifting device is a flat and thin component with only 2.6 mm of thickness, which could be placed in the gap of an LFNED without increasing the thickness. The results show that the proposed components could be easily integrated in an LFNED with the result of resolution enhancement.

Image formation modeling and analysis of near‐eye light field displays

“…Figure shows the basic architecture of an NE‐LFD where an MLA and a microdisplay are placed near an eye. To reconstruct a virtual image point in the 3D virtual image space, several pixels on the microdisplay are so selected to emit light that their chief rays intersect at the intended virtual image point . When the eye accommodation coincides with the virtual image distance, the retina is expected to receive the sharpest image .…”

“…From a signal processing point of view, the minimum sampling interval in the system is the pixel size of the microdisplay, and the interpolation function used for the signal reconstruction on the retina is a combination of a pixel's retinal image and optical point spread function (PSF). Different from regular displays whose interpolation functions are usually within the resolution limit of human eyes, the large ratio of the pixel size to the short focal length of the MLA makes a pixel's retinal image usually larger than the resolution limit; that is, the pixel size plays an important role in NE‐LFDs. To quickly estimate the resolution of an NE‐LFD, currently, several widely used models simplify each factor above into a single value . Usually, in such models, the diffraction‐related resolution β 1 is determined by the Fraunhofer approximation, as a function of the lens aperture p and the wavelength λ : β 1 = p /(1.22 λ ).…”

“…To quickly estimate the resolution of an NE-LFD, currently, several widely used models simplify each factor above into a single value. 20,26,[32][33][34][35] Usually, in such models, the diffraction-related resolution β 1 is determined by the Fraunhofer approximation, as a function of the lens aperture p and the wavelength λ: β 1 = p/(1.22λ).…”

“…To break the resolution limit discussed above, a timemultiplexing method is useful; for example, our previous study 32 used a so-called e-shifting schema to optically shift the light field image for half a pixel in half time domains, leading to an extra resolution increase of approximately 40%.…”

“…To reconstruct a virtual image point in the 3D virtual image space, several pixels on the microdisplay are so selected to emit light that their chief rays intersect at the intended virtual image point. [20][21][22][23]31,32 When the eye accommodation coincides with the virtual image distance, the retina is expected to receive the sharpest image. 18 In this manner, the NE-LFD produces a monocular focus cue matching with the virtual image distance, although the MLA has its native virtual image plane, also known as the central depth plane (CDP).…”