Lightness, illumination, and gradients

The illumination interpretation approach claims that lightness illusions can be explained as misapplications of lightness constancy mechanisms, processes which usually enable veridical extraction of surface reflectance from luminance distributions by discounting illumination. In particular, luminance gradients are thought to provide cues about the interactions of light and surfaces. Several examples of strong lightness illusions are discussed for which explanations based on illumination interpretation can be proposed. In criticisms of this approach, a variety of demonstrations of similarly structured control displays are presented, which involve equivalent lightness effects that cannot readily be accounted for by illumination interpretation mechanisms. Furthermore, a number of known and novel displays are presented that demonstrate effects of gradients on the qualitative appearance of uniform regions. Finally, some simple simulations of neural effects of luminance distributions are discussed.     

Dynamic decorrelation as a unifying principle for explaining a broad range of brightness phenomena

The visual system is highly sensitive to spatial context for encoding luminance patterns. Context sensitivity inspired the proposal of many neural mechanisms for explaining the perception of luminance (brightness). Here we propose a novel computational model for estimating the brightness of many visual illusions. We hypothesize that many aspects of brightness can be explained by a dynamic filtering process that reduces the redundancy in edge representations on the one hand, while non-redundant activity is enhanced on the other. The dynamic filter is learned for each input image and implements context sensitivity. Dynamic filtering is applied to the responses of (model) complex cells in order to build a gain control map. The gain control map then acts on simple cell responses before they are used to create a brightness map via activity propagation. Our approach is successful in predicting many challenging visual illusions, including contrast effects, assimilation, and reverse contrast with the same set of model parameters.     

图形简单几何特征的分辨及其相关错觉的定量研究

本实验用定量的心理物理测量方法,研究了在亮度对比与等亮度颜色对比的条件下,受试者分辨图形简单几何特征的能力以及与这些图形特征有关的错觉效应（Ｚｏｅｌｌｎｅｒ错觉、Ｍｕｌｌｅｒ－Ｌｙｅｒ错觉、Ｐｏｎｚｏ错觉和Ｄｅｌｂｏｅｕｆ错觉）。实验结果表明：受试者在亮度对比与颜色对比的条件下对线段平行度,线段长度与图形面积的分辨能力相同,与这些几何特征有关的错觉幅度也没有显著差异。但是,在等亮度颜色对比条件下,     

The dynamic range of human lightness perception

Natural viewing challenges the visual system with images that have a dynamic range of light intensity (luminance) that can approach 1,000,000:1 and that often exceeds 10,000:1 [1, 2]. The range of perceived surface reflectance (lightness), however, can be well approximated by the Munsell matte neutral scale (N 2.0/ to N 9.5/), consisting of surfaces whose reflectance varies by about 30:1. Thus, the visual system must map a large range of surface luminance onto a much smaller range of surface lightness. We measured this mapping in images with a dynamic range close to that of natural images. We studied simple images that lacked segmentation cues that would indicate multiple regions of illumination. We found a remarkable degree of compression: at a single image location, a stimulus luminance range of 5,905:1 can be mapped onto an extended lightness scale that has a reflectance range of 100:1. We characterized how the luminance-to-lightness mapping changes with stimulus context. Our data rule out theories that predict perceived lightness from luminance ratios or Weber contrast. A mechanistic model connects our data to theories of adaptation and provides insight about how the underlying visual response varies with context.     

Selection of visual information for lightness judgements by eye movements

When judging the lightness of objects, the visual system has to take into account many factors such as shading, scene geometry, occlusions or transparency. The problem then is to estimate global lightness based on a number of local samples that differ in luminance. Here, we show that eye fixations play a prominent role in this selection process. We explored a special case of transparency for which the visual system separates surface reflectance from interfering conditions to generate a layered image representation. Eye movements were recorded while the observers matched the lightness of the layered stimulus. We found that observers did focus their fixations on the target layer, and this sampling strategy affected their lightness perception. The effect of image segmentation on perceived lightness was highly correlated with the fixation strategy and was strongly affected when we manipulated it using a gaze-contingent display. Finally, we disrupted the segmentation process showing that it causally drives the selection strategy. Selection through eye fixations can so serve as a simple heuristic to estimate the target reflectance.     

错觉图形组成成分间的亮度对比和颜色对比对错觉程度的影响

用定量的心理物理测量方法,研究了错觉图形组成成分间的亮度对比和颜色对比方位错觉、长度觉及面积错觉幅度的影响。测试结果表明：与通常的错觉效应相比,当错觉图形组成成分间存在亮度对比或颜色对比（等亮度）时,受试者的错觉程度明显降低;其中,当存在颜色对比时,方位错觉的下降幅度更为显著,达到６９．３％。此外还观察到,在单纯亮度对比条件下,只需１．８％和５．３％的低对比度即可分别产生轮廓和边缘错觉;但在等亮度     

Responses to lightness variations in early human visual cortex

Lightness is the apparent reflectance of a surface, and it depends not only on the actual luminance of the surface but also on the context in which the surface is viewed [1-10]. The cortical mechanisms of lightness processing are largely unknown, and the role of early cortical areas is still a matter of debate [11-17]. We studied the cortical responses to lightness variations in early stages of the human visual system with functional magnetic resonance imaging (fMRI) while observers were performing a demanding fixation task. The set of dynamically presented visual stimuli included the rectangular version of the classic Craik-O'Brien stimulus [3, 18, 19] and a variant that led to a weaker lightness effect, as well as a pattern with actual luminance variations. We found that the cortical activity in retinotopic areas, including the primary visual cortex (V1), is correlated with context-dependent lightness variations.     

Visual perception: more than meets the eye

A recent study shows that objects changing in colour, luminance, size or shape appear to stop changing when they move. These and other compelling illusions provide tantalizing clues about the mechanisms and limitations of object analysis.     

Distinct mechanisms mediate visual detection and identification

A core organizing principle for studies of the brain is that distinct neural pathways mediate distinct behavioral tasks [1, 2]. When two related tasks are mediated by a common pathway, studies of one are likely to generalize to the other. Here, we test whether performance on two laboratory tasks that model object detection and identification are mediated by common mechanisms of visual adaptation. Although both tasks rely on the luminance pattern in images, their demands on visual processing are quite different. Object detection requires discriminating image luminance differences associated with the light reflected from adjacent objects. To encode these differences reliably, neurons adapt their limited dynamic range to prevailing viewing conditions [3-6]. Object identification, on the other hand, demands a fixed response to light reflected from an object independent of illumination [7]. We compared performance in discrimination and identification tasks for simulated surfaces. In striking contrast to studies with less structured contexts, we found clear evidence that distinct processes mediate judgments in the two tasks. These results challenge models that account for perceived lightness entirely through the action of image-encoding mechanisms.     

Visual perception: lightness in a high-dynamic-range world

How the visual system detects surface reflectance - lightness - has puzzled researchers for centuries. A new study of contexts that capture the characteristic luminance range of natural scenes rules out a long-standing theory of lightness perception.     

Reinterpreting Behavioral Receptive Fields: Lightness Induction Alters Visually Completed Shape

Background

A classification image (CI) technique has shown that static luminance noise near visually completed contours affects the discrimination of fat and thin Kanizsa shapes. These influential noise regions were proposed to reveal “behavioral receptive fields” of completed contours–the same regions to which early cortical cells respond in neurophysiological studies of contour completion. Here, we hypothesized that 1) influential noise regions correspond to the surfaces that distinguish fat and thin shapes (hereafter, key regions); and 2) key region noise biases a “fat” response to the extent that its contrast polarity (lighter or darker than background) matches the shape''s filled-in surface color.

Results

To test our hypothesis, we had observers discriminate fat and thin noise-embedded rectangles that were defined by either illusory or luminance-defined contours (Experiment 1). Surrounding elements (“inducers”) caused the shapes to appear either lighter or darker than the background–a process sometimes referred to as lightness induction. For both illusory and luminance-defined rectangles, key region noise biased a fat response to the extent that its contrast polarity (light or dark) matched the induced surface color. When lightness induction was minimized, luminance noise had no consistent influence on shape discrimination. This pattern arose when pixels immediately adjacent to the discriminated boundaries were excluded from the analysis (Experiment 2) and also when the noise was restricted to the key regions so that the noise never overlapped with the physically visible edges (Experiment 3). The lightness effects did not occur in the absence of enclosing boundaries (Experiment 4).

Conclusions

Under noisy conditions, lightness induction alters visually completed shape. Moreover, behavioral receptive fields derived in CI studies do not correspond to contours per se but to filled-in surface regions contained by those contours. The relevance of lightness to two-dimensional shape completion supplies a new constraint for models of object perception.     

Color and lightness constancy in different perceptual tasks

Color and lightness constancy with respect to changing illumination was studied with three different perceptual tasks: ranking of colored papers according (1) to their lightness and (2) to their chromatic similarity in photopic, mesopic, and scotopic states of adaptation, and (3) recognition of remembered colored papers after changes of illumination in photopic vision. Constancy was found in the second task, only. Excitations of light receptors and luminance channels were computed to simulate the empirical rank orders. Results of the first task can be predicted with the hypothesis that luminance channels are activated, if lightness is asked for. Sequences arranged with respect to chromatic similarity were found independent of the illuminant spectra, even if the calculated rank orders of cone excitation were changed in the altered illumination. Received: 4 October 1997 / Accepted in revised form: 26 August 1998     

The knowledge used in vision and where it comes from.

Knowledge is often thought to be something brought from outside to act upon the visual messages received from the eye in a ''top-down'' fashion, but this is a misleadingly narrow view. First, the visual system is a multilevel heterarchy with connections acting in all directions so it has no ''top''; and second, knowledge is provided through innately determined structure and by analysis of the redundancy in sensory messages themselves, as well as from outside. This paper gives evidence about mechanisms analysing sensory redundancy in biological vision. Automatic gain controls for luminance and contrast depend upon feedback from the input, and there are strong indications that the autocorrelation function, and other associations between input variables, affect the contrast sensitivity function and our subjective experience of the world. The associative structure of sensory message can provide much knowledge about the world we live in, and neural mechanisms that discount established associative structure in the input messages by recoding them can improve survival by making new structure more easily detectable. These mechanisms may be responsible for illusions, such as those produced by a concave face-mask, that are classically attributed to top-down influences.     

Natural image statistics mediate brightness 'filling in'

Although the human visual system can accurately estimate the reflectance (or lightness) of surfaces under enormous variations in illumination, two equiluminant grey regions can be induced to appear quite different simply by placing a light-dark luminance transition between them. This illusion, the Craik-Cornsweet-O'Brien (CCOB) effect, has been taken as evidence for a low-level 'filling-in' mechanism subserving lightness perception. Here, we present evidence that the mechanism responsible for the CCOB effect operates not via propagation of a neural signal across space but by amplification of the low spatial frequency (SF) structure of the image. We develop a simple computational model that relies on the statistics of natural scenes actively to reconstruct the image that is most likely to have caused an observed series of responses across SF channels. This principle is tested psychophysically by deriving classification images (CIs) for subjects' discrimination of the contrast polarity of CCOB stimuli masked with noise. CIs resemble 'filled-in' stimuli; i.e. observers rely on portions of the stimuli that contain no information per se but that correspond closely to the reported perceptual completion. As predicted by the model, the filling-in process is contingent on the presence of appropriate low SF structure.     

A neural model of surface perception: lightness, anchoring, and filling-in

A neural model is proposed of how the visual system processes natural images under variable illumination conditions to generate surface lightness percepts. Previous models clarify how the brain can compute relative contrast. The anchored Filling-In Lightness Model (aFILM) clarifies how the brain 'anchors' lightness percepts to determine an absolute lightness scale that uses the full dynamic range of neurons. The model quantitatively simulates lightness anchoring properties (Articulation, Insulation, Configuration, Area Effect) and other lightness data (discounting the illuminant, the double brilliant illusion, lightness constancy and contrast, Mondrian contrast constancy, Craik-O'Brien-Cornsweet illusion). The model clarifies how retinal processing stages achieve light adaptation and spatial contrast adaptation, and how cortical processing stages fill-in surface lightness using long-range horizontal connections that are gated by boundary signals. The new filling-in mechanism runs 1000 times faster than diffusion mechanisms of previous filling-in models.     

Disruptive contrast in animal camouflage

Camouflage typically involves colour patterns that match the background. However, it has been argued that concealment may be achieved by strategic use of apparently conspicuous markings. Recent evidence supports the theory that the presence of contrasting patterns placed peripherally on an animal's body (disruptive coloration) provides survival advantages. However, no study has tested a key prediction from the early literature that disruptive coloration is effective even when some colour patches do not match the background and have a high contrast with both the background and adjacent pattern elements (disruptive contrast). We test this counter-intuitive idea that conspicuous patterns might aid concealment, using artificial moth-like targets with pattern elements designed to match or mismatch the average luminance (lightness) of the trees on which they were placed. Disruptive coloration was less effective when some pattern elements did not match the background luminance. However, even non-background-matching disruptive patterns reduced predation relative to equivalent non-disruptive patterns or to unpatterned controls. Therefore, concealment may still be achieved even when an animal possesses markings not found in the background. Disruptive coloration may allow animals to exploit backgrounds on which they are not perfectly matched, and to possess conspicuous markings while still retaining a degree of camouflage.     

Brightness versus apparent contrast. 2: Large-field asymmetry.

Experiments were performed on the quasi-static perception of brightness and of apparent contrast of a foveal 1-deg disk, presented either as a luminance increment or decrement against a 300 cd.m-2 background. Results suggest that the perceptual attributes of brightness and apparent (or subjective) contrast should be distinguished. For an equal brightness difference with respect to the background, luminance increments are more effective than decrements. For an equal apparent contrast it is found that increments and decrements, up to 100 cd.m-2, are about equally effective; for higher values luminance decrements are more effective. Brightness increments and decrements can both be described by a Stevens power function of the respective luminance increments and decrements. Apparent contrast can, apart from applying a usual luminance contrast formula, also be described as a power function of the luminance difference with the background.     

An account of responses of spectrally opponent neurons in macaque lateral geniculate nucleus to successive contrast

Coloured surfaces in the normal environment may be brighter or dimmer than the mean adaptation level. Changes in the firing rate of cells of the parvocellular layers of macaque lateral geniculate nucleus were studied with such stimuli; chromatic mixtures briefly replaced a white adaptation field. This paradigm is therefore one of successive contrast. Families of intensity-response curves for different wavelengths were measured. When taking sections at different luminance ratios through these families of curves, strongly opponent cells displayed spectrally selective responses at low luminance ratios, while weakly opponent cells had higher chromatic thresholds and responded well to stimuli at higher luminance ratios, brighter than the adaptation field. Strength of cone opponency, defined as the weight of the inhibitory cone mechanism relative to the excitatory one, was thus related to the range of intensity in which cells appeared to operate most effectively. S-cone inputs, as tested with lights lying along tritanopic confusion lines, could either be excitatory or inhibitory. Families of curves for different wavelengths can be simulated mathematically for a given cell by a simple model by using known cone absorption spectra. Hyperbolic response functions relate cone absorption to the output signals of the three cone mechanisms, which are assumed to interact linearly. Parameters from the simulation provided estimates of strength of cone opponency and cone sensitivity which were shown to be continuously distributed. Cell activity can be related to cone excitation in a trichromatic colour space with the help of the model, to give an indication of suprathreshold coding of colour and lightness.     

Color and luminance contrasts attract independent attention

Paying attention can improve vision in many ways, including some very basic functions such as contrast discrimination, a task that probably reflects very early levels of visual processing. Electrophysiological, psychophysical, and imaging studies on humans as well as single recordings in monkey show that attention can modulate the neuronal response at an early stage of visual processing, probably by acting on the response gain. Here, we measure incremental contrast thresholds for luminance and color stimuli to derive the contrast response of early neural mechanisms and their modulation by attention. We show that, for both cases, attention improves contrast discrimination, probably by multiplicatively increasing the gain of the neuronal response to contrast. However, the effects of attention are highly specific to the visual modality: concurrent attention to a competing luminance, but not chromatic pattern, greatly impedes luminance contrast discrimination; and attending to a competing chromatic, but not luminance, task impedes color contrast discrimination. Thus, the effects of attention are highly modality specific, implying separate attentional resources for different fundamental visual attributes at early stages of visual processing.     

Humans are not fooled by size illusions in attractiveness judgements

Could signallers use size contrast illusions to dishonestly exaggerate their attractiveness to potential mates? Using composite photographs of women from three body mass index (BMI) categories designed to simulate small groups, we show that target women of medium size are judged as thinner when surrounded by larger women than when surrounded by thinner women. However, attractiveness judgements of the same target women were unaffected by this illusory change in BMI, despite small true differences in the BMIs of the target women themselves producing strong effects on attractiveness. Thus, in the context of mate choice decisions, the honesty of female body size as a signal of mate quality appears to have been maintained by the evolution of assessment strategies that are immune to size contrast illusions. Our results suggest that receiver psychology is more flexible than previously assumed, and that illusions are unlikely to drive the evolution of exploitative neighbour choice in human sexual displays.