Monocular vs. stereo cues

Figure 6.1: This painting uses a monocular depth cue called a texture gradient to enhance depth perception: The bricks become smaller and thinner as the depth increases. Other cues arise from perspective projection, including height in the visual field and retinal image size. (``Paris Street, Rainy Day,'' Gustave Caillebotte, 1877. Art Institute of Chicago.)

Figure 6.2: Even simple line drawings provide significant cues. (a) The Ponzo illusion: The upper yellow bar appears to be longer, but both are the same length. (b) The Müller-Lyer illusion: The lower horizontal segment appears to be shorter than the one above, but they are the same length.
...erlyer.eps,width=2.4truein} \\
(a) & (b)

A piece of information that is derived from sensory stimulation and is relevant for perception is called a sensory cue or simply a cue. In this section, we consider only depth cues, which contribute toward depth perception. If a depth cue is derived from the photoreceptors or movements of a single eye, then it is called a monocular depth cue. If both eyes are required, then it is a stereo depth cue. There are many more monocular depth cues than stereo, which explains why we are able to infer so much depth information from a single photograph. Figure 6.1 shows an example. The illusions in Figure 6.2 show that even simple line drawings are enough to provide strong cues. Interestingly, the cues used by humans also work in computer vision algorithms to extract depth information from images [318].

Steven M LaValle 2016-12-31