Human factor study on the crosstalk of multiview autostereoscopic displays

Stereoscopic depth perception has been analyzed in many laboratory experiments since Wheatstone's (1838) discovery that disparity is a sufficient and compelling stimulus for the perception of depth with mirror-type stereo displays. In this paper, mirror-type stereo displays were used as the instrument to simulate the 3D image in the human factor experiment. It can be used to simulate the 9 view 3D display by image processing method with different multi-view crosstalk levels measured from luminance measurement device. The disparity of multi-view images to form stereopsis with depth perception is decided by the 9-view autostereoscopic 3D display that subject can properly fuse the image to get the proper visual depth. Computer graphic method applied for multi-view content rendering with shooting distance of 70 cm for each virtual camera. The distance between cameras is 5.6 cm with parallel capture to simulate the images accepted by human eyes. The experimental design was used for testing subjective evaluations based on the questionnaire, and ANOVA methods were used for analysis. Experimental variables of this human factor study for multi-view 3D display are five levels of crosstalk distribution from measured data, with or without shadow effects and perspective line shown within tested images. In addition, the result of acceptable system crosstalk level for multi-view stereoscopic display is between Level 4.7 and Level 5.9 in average for the four tested images.