Geometric Distortions Related Visually Induced Motion Sickness Prediction And Psychovisual Modulation Technology In Stereoscopic Displays

Stereoscopic(i.e.,stereoscopic 3D or S3D)displays are being applied to a number of different areas,such as 3D movies,scientific visualization,medical imaging,3D games,minimally invasive surgery,and virtual/augmented reality.Compared to conventional 2D displays,S3D displays provide extra depth information by sending two slightly different perspective images to the viewer's left and right eyes separately.The ability to provide extra depth information makes the user's 3D viewing experience more immersive or depth prediction more accurate and precise.However,people exposed to S3D displays report motion-sickness-like symptoms,such as nausea,headache,salivation,sweating,dizziness,and vertigo.Since the discomfort symptoms are invoked by visual environments alone without the viewer's physical motions,this discomfort is called visually induced motion sickness(VIMS).First,geometric distortions in S3D imaging are studied.Motion in a distorted S3D space may cause VIMS.S3D imaging chain includes the capture of original 3D scenes using two cameras,the display of 3D content on S3D displays,and users' S3D viewing.The mismatches among the capture,display,and viewing result in geometric distortions in S3D.Six types of potential mismatches are considered,including camera separation vs.eye separation,camera field of view(FOV)vs.screen FOV,camera convergence distance vs.screen distance from the observer,head position vs.display image position,head orientation vs.display image orientation,and camera image plane vs.display image plane.The geometric model in this thesis can predict the geometric distortions caused by each kind of mismatch.The model facilitates insights into the various distortions and leads us to a method whereby the user can correct geometric distortions by adjusting some variables to alter the effects of others.Moreover,the geometric distortions are analyzed from the first person's point of view.Geometric distortions result in conflicts between monocular and binocular depth cues,resealing in VIMS in S3D viewing.Second,this thesis proposes a method to quantify VIMS using VIMS level rating during the S3D viewing.In order to solve the VIMS problem in S3D,we need to first be able to measure it.A variety of measurement techniques related to VIMS were used in the previous studies,either objectively or subjectively.The simulator sickness questionnaire(SSQ)or postural stability measures were two commonly used tools.But they only measure the changes between pre-and post-experiment.A novel VIMS susceptibility measure from VIMS level rating during the experiment is introduced by combining measures of the subject's 'sensitivity' and 'endurance'to the VIMS.The proposed VIMS susceptibility measure was tested for various VIMS inducing conditions,and showed the effectiveness in conducting both between-subjects and within-subject comparisons for different VIMS conditions.Another contribution of this thesis is a display technology called psychovisual modulation technology,which is the extension of stereoscopic display technology and was first proposed by Wu and Zhai(2013).S3D displays using spatial or temporal multiplexing techniques have high spatial or temporal resolutions so that a single display is able to present two views for users.The high capability of displays with high spatial or temporal resolutions can be explored for other usages instead of simply multiplexing two views for the left and right eyes.For example,the display can present a hidden/personal view for users with a pair of specific glasses,meanwhile,present a shared view for users without the glasses.The spatial or temporal resolutions of some displays are beyond the spatial and temporal discrimination thresholds of human visual systems(HVS).Thus,the psychovisual redundancy of the display provides extra capacity to HVS.Based on the work by Wu and Zhai(2013),additional two extra constraints are considered:the shared view without using viewing devices should be considered and the modulation weights should be sparse due to the material property of LC.We also propose a modified version of hierarchical alternating least squares(HALS)algorithms to solve the constrained NMF problem.Moreover,a spatially-weighted NMF is applied to psychovisual modulation problems to improve the visual quality of display images.At last,several applications are developed based on the psychovisual modulation technol-ogy.There are mainly two ways to utilize the psychovisual redundancy of displays.First,the psychovisual redundancy of display device can be used to generate multiple visual percepts for different viewers,such as information security displays and dual-view medical image visual-ization.Second,the redundancy can be used for data transmission to computing devices(i.e.,visible light communication from display devices to computing devices).Thus,a unidirectional communication channel from display device to computing device can be widely used in many fields,such as anti-piracy displays,invisible QR codes displays.In summary,this thesis studies on S3D displays and can mainly be separated into two parts:VIMS caused by geometric distortions and psychovisual modulation technology.The goal of the VIMS study is to reduce the discomfort for S3D viewing so that more people can enjoy the stereoscopic viewing.The study on psychovisual modulation extends the usage of S3D displays so that the extra capability of S3D displays can be fully explored.Stereoscopic display technology is a special case of the psychovisual modulation technology.In short,this thesis works on improving stereoscopic display technology which is what S3D displays are originally designed for and then generalizing to psychovisual modulation technology for other usages.