The human visual system's assumption that light comes from above is weak

Abstract

Every biological or artificial visual system faces the problem that images are highly ambiguous, in the sense that every image depicts an infinite number of possible 3D arrangements of shapes, surface colors, and light sources. When estimating 3D shape from shading, the human visual system partly resolves this ambiguity by relying on the light-from-above prior, an assumption that light comes from overhead. However, light comes from overhead only on average, and most images contain visual information that contradicts the light-from-above prior, such as shadows indicating oblique lighting. How does the human visual system perceive 3D shape when there are contradictions between what it assumes and what it sees? Here we show that the visual system combines the light-from-above prior with visual lighting cues using an efficient statistical strategy that assigns a weight to the prior and to the cues and finds a maximum-likelihood lighting direction estimate that is a compromise between the two. The prior receives surprisingly little weight and can be overridden by lighting cues that are barely perceptible. Thus, the light-from-above prior plays a much more limited role in shape perception than previously thought, and instead human vision relies heavily on lighting cues to recover 3D shape. These findings also support the notion that the visual system efficiently integrates priors with cues to solve the difficult problem of recovering 3D shape from 2D images.

The perception of shape from shading in a new light

“…The present findings provide converging support with recent behavioural evidence that 174 lighting cues, rather than a 'light from above' bias, are the primary determinant of the perception 175 of shape from shading [9]. This results reported here provide further evidence that the perception 176 of shape from shading, like other aspects of visual perception [18], arises from interactions with 177 the natural world rather than internal biases that give rise to illusory perception.…”

“…Given the evidence presented by a number of studies, summarized and tested 97 by Morgenstern and colleagues [9], it should be well known now that the original finding of a 98 light from overhead bias not the whole story. Despite such evidence to the contrary, the light 99 from overhead view is still often cited [5][6][7].…”

“…The 'single light from above' hypothesis [1,4] in the perception of shape from shading [9]. Morgenstern and colleagues presented lighting cues 61 that could influence shape perception in either accordance with an objects context in the image 62 or instead with the light from above prior.…”

Investigation and remediation of false topographic perception phenomena observed on Chang’E-1 lunar imagery

“…Human brain perceives that light illuminating source is from northern hemisphere (front) rather than from southern hemisphere (behind). Human vision relies heavily on lighting cues to recover 3D shape (Morgenstern et al, 2011). When estimating 3D shapes from shading, the human visual system resolves this ambiguity by relying on the prior knowledge of illumination from light-from-front.…”

How reliable is perception?

“…As recently demonstrated by Witzel et al, manipulating expectations about whether the dress is illuminated by daylight or tungsten light influences the colors that people see (Witzel, Racey, & O'Regan, in press). surfaces are often lit from above can help us infer depth when the direction of the illuminant is ambiguous (Adams, 2007); this "light from above" is flexibly updated to reflect information about the light source as it becomes available (e.g., Morgenstern, Murray, & Harris, 2011). Knowledge in one modality should influence other modalities when the modalities are mutually informative.…”