The Simulation Of Gaze Effect Based On Eye Physiology

In recent years, gaze-contingent rendering technique on human visual perception has been a hot research focus. Applications include flight and driving simulators, virtual reality, infrared and indirect vision, remote piloting, robotics and automation, teleoperation, and telemedicine; image transmission and retrieval, and video teleconferencing. This paper process two gaze-contingent rendering techniques. The former is a nonlinear Mipmap interpolation technique based on image post processing. The later is a based-model non-uniform sampling rendering technique.In chapter one, we introduce the crucial significance of gaze-contingent rendering technique, at the same time, reviewing the related work in the area of gaze-contingent rendering technique, and briefly describing the main aspect of the research of this paper.The second chapter describes the related basic knowledge of physiological and psychological theory concerning the human visual system, which provides a basis for our algorithm design and establishment.In chapter three, we suggest an efficient gaze-contingent rendering algorithm based on Mipmap Pyramide. At first the spatial-variant Gaussian kernel is calculated according to critical frequency; then we relate it to corresponding Mipmap level by using pixel-wise method of Root-Mean-Square-Error (RMSE) to find a best match; due to the nonlinear Mipmap interpolation under Bilateral Filtering scheme, we obtain the final color for display to simulate the gaze effect of human eye in realtime.In chapter four, we suggest a non-uniform sampling gaze-contingent rendering algorithm. At the first, an thin len system is used to simulate human eye optical system; then,In order to avoid testing lens rays against an actual scene and, instead, derive a layered image-based representation via depth peeling from the lens'center. we generate image samples which is consistent with the human eye perceived based on visual acuity equation; we use gaussian weighted interpolation to calculate the color of r non-sample points; at last, due to the realtime scheme, we use CPU technology to accelerate the algorithmThis work is summarized at the last chapter, as well as the prospect of future research.