Binocular disparity

Fundamentals of Liquid Crystal Devices, Second Edition. Deng-Ke Yang and Shin-Tson Wu. 2015 John Wiley Sons, Ltd. Published 2015 by John Wiley Sons, Ltd. 

---

Both types of stereo presentation mimic the appearance of objects to our two eyes, which produce images on the retina of objects seen from two slightly different viewpoints. The two images are rotated about a vertical axis located at the current focal point of the eyes. From the difference between the two images, called binocular disparity, we obtain information about the relative depth of objects in our field of view. For viewers with normal vision,two pic-... 

When working by visual information, it is necessary to grasp what you have in the outside world at sight, and a result of recognition must be reflected by an action. A two-dimensional moving point with one camera mounted on a robot was chased [1-3]. But, two cameras are necessary in the case of three dimensions. Binocular disparity of a base-line stereo that both optical axes are parallel is usually used [4]. Gaze control is a method to determine an object on a intersection point of optical axes of both cameras and makes it possible to always cateh an observation objeet at the center of a field of vision [5, 6]. This method has coneentration of gaze... 

Figure 16.1 Illustration of binocular disparity an image observed from (a) left eye, and (b) right eye.

When a human head moves sidewise, the pupil location changes so that the viewing direction changes accordingly. The closer objects appear to move faster across the field of view than those further away. This effect is, in principle, similar to binocular disparity. The former is a result of the temporal change of viewing points, and the latter is a result of spatially separated... 

Binocular disparity and convergence require the involvement of both eyes, while motion parallax and accommodation could be observed eveu with a single eye. For natural 3-D objects, all the four major depth cues meutioned above should be present at the same lime. Various 3-D display technologies employ at least one of the four major depth cues to geuerate 3-D depth sensation. The more consistent the depth cues are, the more realistic and natural a 3-D image appears. 

A stereoscopic display requires viewers to wear special glasses in order to see two sKghtly different 2D images in two different eyes. The 2D images are integrated by human brain to generate 3D depth perception. Apparently, the primary depth cue of stereoscopic displays is binocular disparity. Several types of stereoscopic displays have been developed, as discussed below. 

Fig. 13 Angular binocular disparity

Distance estimation from binocular disparity requires an offset in the relative retinal projection position of the same feature in both eyes. Figure 13 illustrates how objects at different depths manifest as retinal images with different disparities d on. The distance of fixation is p in which the relative disparity is zero (base width of the stereo system b ra 65 mm and the distance to the point of fixation P as well as on the object distance o. To calculate the disparity shear angle y has to be calculated. 

These limits are derived from the assumption of a perfect ability of the neurons in the visual cortex to process and correlate the features in each hemisphere to compute disparity. Howard and Rogers [49] have confirmed the minimum neuronal stereo processing capability of a resolution down to 0.6 arcmin. However, studies by Glennester [50] have shown that the maximum detectable disparity is 20 arcmin which equals about 33 photoreceptor spacings. Hence, the depth cue of binocularity is specialized for short distances <5-10 m. Similar results to these theoretical considerations have been verified by experimental approaches from Blakemore [51]. 

---