Experimental Study On Estimating Gamma Value Of VR Display Using Human Vision

With the development of virtual reality(VR)technology,3D head-mounted display(HMD)are widely used.However,due to the influence of hardware and other factors,the gamma values of the two displays cannot be guaranteed to be the same,resulting in different colors.When the color difference between the left and right screens of 3D HMDs is too large,users will have binocular competition,or even visual discomfort and fatigue symptoms,which seriously restricts the development of VR industry.Gamma correction on the left and right display screens of 3D HMDs can effectively reduce the size of binocular color difference.Traditional methods can only be carried out with the help of professional calibration instruments,which are expensive,time-consuming and laborious,and not suitable for ordinary users.Based on fringe matching method,this paper designs 3D HMD software for human vision estimation,invites subjects to perform estimation experiments,and studies human vision estimation and 3D HMD gamma value from two aspects: visual estimation of3 D HMD gamma value and comparison of estimated value under different parallax.Specific contents include:1.The experimental software of objective measurement of 3D HMD display gamma value was developed with Unity,and the spectrometer PR-715 was used to measure 10 3D HMDs.The results showed that there was a general difference in the left and right display of 3D HMD gamma value,with an average difference of 0.161.2.A set of gamma estimation software based on human vision observation was developed based on fringe matching method.The software was used to estimate the gamma value of VR display,and a piecewise fitting method was proposed.The average error of gamma value was 0.057 compared with the experimental results.The resulting color error is approximately 2.04 CIELAB color difference units.The analysis results show that it is feasible to estimate the gamma value of 3D HMD by human vision.3.Experiments were carried out on whether parallax clues would affect gamma estimation.The experimental results showed that when parallax was-40 arcs,human vision had the highest accuracy in estimating gamma value of VR display,and the estimation error was 0.045.When parallax was divided into-40 arcs as the baseline,the error would increase regardless of the virtual depth from near to far or from far to near.This indicates that the parallax clue has an effect on the visual gamma value estimation of the subjects in VR display.