Surface Reflectance Recognition and Real-World Illumination Statistics

Humans distinguish materials such as metal, plastic, and paper effortlessly at a glance. Traditional computer vision systems cannot solve this problem at all. Recognizing surface reflectance properties from a single photograph is difficult because the observed image depends heavily on the amount of light incident from every direction. A mirrored sphere, for example, produces a different image in every environment. To make matters worse, two surfaces with different reflectance properties could produce identical images. The mirrored sphere simply reflects its surroundings, so in the right artificial setting, it could mimic the appearance of a matte ping-pong ball. Yet, humans possess an intuitive sense of what materials typically "look like" in the real world. This thesis develops computational algorithms with a similar ability to recognize reflectance properties from photographs under unknown, real-world illumination conditions. Real-world illumination is complex, with light typically incident on a surface from every direction. We find, however, that real-world illumination patterns are not arbitrary. They exhibit highly predictable spatial structure, which we describe largely in the wavelet domain. Although they differ in several respects from the typical photographs, illumination patterns share much of the regularity described in the natural image statistics literature. These properties of real-world illumination lead to predictable image statistics for a surface with given reflectance properties. We construct a system that classifies a surface according to its reflectance from a single photograph under unknown illuminination. Our algorithm learns relationships between surface reflectance and certain statistics computed from the observed image. Like the human visual system, we solve the otherwise underconstrained inverse problem of reflectance estimation by taking advantage of the statistical regularity of illumination. For surfaces with homogeneous reflectance properties and known geometry, our system rivals human performance.

View-Based Strategies for 3D Object Recognition

A persistent issue of debate in the area of 3D object recognition concerns the nature of the experientially acquired object models in the primate visual system. One prominent proposal in this regard has expounded the use of object centered models, such as representations of the objects' 3D structures in a coordinate frame independent of the viewing parameters [Marr and Nishihara, 1978]. In contrast to this is another proposal which suggests that the viewing parameters encountered during the learning phase might be inextricably linked to subsequent performance on a recognition task [Tarr and Pinker, 1989; Poggio and Edelman, 1990]. The 'object model', according to this idea, is simply a collection of the sample views encountered during training. Given that object centered recognition strategies have the attractive feature of leading to viewpoint independence, they have garnered much of the research effort in the field of computational vision. Furthermore, since human recognition performance seems remarkably robust in the face of imaging variations [Ellis et al., 1989], it has often been implicitly assumed that the visual system employs an object centered strategy. In the present study we examine this assumption more closely. Our experimental results with a class of novel 3D structures strongly suggest the use of a view-based strategy by the human visual system even when it has the opportunity of constructing and using object-centered models. In fact, for our chosen class of objects, the results seem to support a stronger claim: 3D object recognition is 2D view-based.

Reciprocal Interactions Between Motion and Form Perception

The processes underlying the perceptual analysis of visual form are believed to have minimal interaction with those subserving the perception of visual motion (Livingstone and Hubel, 1987; Victor and Conte, 1990). Recent reports of functionally and anatomically segregated parallel streams in the primate visual cortex seem to support this hypothesis (Ungerlieder and Mishkin, 1982; VanEssen and Maunsell, 1983; Shipp and Zeki, 1985; Zeki and Shipp, 1988; De Yoe et al., 1994). Here we present perceptual evidence that is at odds with this view and instead suggests strong symmetric interactions between the form and motion processes. In one direction, we show that the introduction of specific static figural elements, say 'F', in a simple motion sequence biases an observer to perceive a particular motion field, say 'M'. In the reverse direction, the imposition of the same motion field 'M' on the original sequence leads the observer to perceive illusory static figural elements 'F'. A specific implication of these findings concerns the possible existence of (what we call) motion end-stopped units in the primate visual system. Such units might constitute part of a mechanism for signalling subjective occluding contours based on motion-field discontinuities.

Detecting Faces in Impoverished Images

The ability to detect faces in images is of critical ecological significance. It is a pre-requisite for other important face perception tasks such as person identification, gender classification and affect analysis. Here we address the question of how the visual system classifies images into face and non-face patterns. We focus on face detection in impoverished images, which allow us to explore information thresholds required for different levels of performance. Our experimental results provide lower bounds on image resolution needed for reliable discrimination between face and non-face patterns and help characterize the nature of facial representations used by the visual system under degraded viewing conditions. Specifically, they enable an evaluation of the contribution of luminance contrast, image orientation and local context on face-detection performance.

Perceiving Illumination Inconsistencies in Scenes

The human visual system is adept at detecting and encoding statistical regularities in its spatio-temporal environment. Here we report an unexpected failure of this ability in the context of perceiving inconsistencies in illumination distributions across a scene. Contrary to predictions from previous studies [Enns and Rensink, 1990; Sun and Perona, 1996a, 1996b, 1997], we find that the visual system displays a remarkable lack of sensitivity to illumination inconsistencies, both in experimental stimuli and in images of real scenes. Our results allow us to draw inferences regarding how the visual system encodes illumination distributions across scenes. Specifically, they suggest that the visual system does not verify the global consistency of locally derived estimates of illumination direction.

Slow and Smooth: A Bayesian Theory for the Combination of Local Motion Signals in Human Vision

In order to estimate the motion of an object, the visual system needs to combine multiple local measurements, each of which carries some degree of ambiguity. We present a model of motion perception whereby measurements from different image regions are combined according to a Bayesian estimator --- the estimated motion maximizes the posterior probability assuming a prior favoring slow and smooth velocities. In reviewing a large number of previously published phenomena we find that the Bayesian estimator predicts a wide range of psychophysical results. This suggests that the seemingly complex set of illusions arise from a single computational strategy that is optimal under reasonable assumptions.

Relative Contributions of Internal and External Features to Face Recognition

The central challenge in face recognition lies in understanding the role different facial features play in our judgments of identity. Notable in this regard are the relative contributions of the internal (eyes, nose and mouth) and external (hair and jaw-line) features. Past studies that have investigated this issue have typically used high-resolution images or good-quality line drawings as facial stimuli. The results obtained are therefore most relevant for understanding the identification of faces at close range. However, given that real-world viewing conditions are rarely optimal, it is also important to know how image degradations, such as loss of resolution caused by large viewing distances, influence our ability to use internal and external features. Here, we report experiments designed to address this issue. Our data characterize how the relative contributions of internal and external features change as a function of image resolution. While we replicated results of previous studies that have shown internal features of familiar faces to be more useful for recognition than external features at high resolution, we found that the two feature sets reverse in importance as resolution decreases. These results suggest that the visual system uses a highly non-linear cue-fusion strategy in combining internal and external features along the dimension of image resolution and that the configural cues that relate the two feature sets play an important role in judgments of facial identity.

Face processing in humans is compatible with a simple shape-based model of vision

Understanding how the human visual system recognizes objects is one of the key challenges in neuroscience. Inspired by a large body of physiological evidence (Felleman and Van Essen, 1991; Hubel and Wiesel, 1962; Livingstone and Hubel, 1988; Tso et al., 2001; Zeki, 1993), a general class of recognition models has emerged which is based on a hierarchical organization of visual processing, with succeeding stages being sensitive to image features of increasing complexity (Hummel and Biederman, 1992; Riesenhuber and Poggio, 1999; Selfridge, 1959). However, these models appear to be incompatible with some well-known psychophysical results. Prominent among these are experiments investigating recognition impairments caused by vertical inversion of images, especially those of faces. It has been reported that faces that differ "featurally" are much easier to distinguish when inverted than those that differ "configurally" (Freire et al., 2000; Le Grand et al., 2001; Mondloch et al., 2002) ??finding that is difficult to reconcile with the aforementioned models. Here we show that after controlling for subjects' expectations, there is no difference between "featurally" and "configurally" transformed faces in terms of inversion effect. This result reinforces the plausibility of simple hierarchical models of object representation and recognition in cortex.

The Perception of Subjective Surfaces

It is proposed that subjective contours are an artifact of the perception of natural three-dimensional surfaces. A recent theory of surface interpolation implies that "subjective surfaces" are constructed in the visual system by interpolation between three-dimensional values arising from interpretation of a variety of surface cues. We show that subjective surfaces can take any form, including singly and doubly curved surfaces, as well as the commonly discussed fronto-parallel planes. In addition, it is necessary in the context of computational vision to make explicit the discontinuities, both in depth and in surface orientation, in the surfaces constructed by interpolation. It is proposed that subjective surfaces and subjective contours are demonstrated. The role played by figure completion and enhanced brightness contrast in the determination of subjective surfaces is discussed. All considerations of surface perception apply equally to subjective surfaces.

Biologically motivated model for object detection and identification in real-world scenes

The classical computer vision methods can only weakly emulate some of the multi-level parallelisms in signal processing and information sharing that takes place in different parts of the primates’ visual system thus enabling it to accomplish many diverse functions of visual perception. One of the main functions of the primates’ vision is to detect and recognise objects in natural scenes despite all the linear and non-linear variations of the objects and their environment. The superior performance of the primates’ visual system compared to what machine vision systems have been able to achieve to date, motivates scientists and researchers to further explore this area in pursuit of more efficient vision systems inspired by natural models. In this paper building blocks for a hierarchical efficient object recognition model are proposed. Incorporating the attention-based processing would lead to a system that will process the visual data in a non-linear way focusing only on the regions of interest and hence reducing the time to achieve real-time performance. Further, it is suggested to modify the visual cortex model for recognizing objects by adding non-linearities in the ventral path consistent with earlier discoveries as reported by researchers in the neuro-physiology of vision.

View-based approaches to spatial representation in human vision

In an immersive virtual environment, observers fail to notice the expansion of a room around them and consequently make gross errors when comparing the size of objects. This result is difficult to explain if the visual system continuously generates a 3-D model of the scene based on known baseline information from interocular separation or proprioception as the observer walks. An alternative is that observers use view-based methods to guide their actions and to represent the spatial layout of the scene. In this case, they may have an expectation of the images they will receive but be insensitive to the rate at which images arrive as they walk. We describe the way in which the eye movement strategy of animals simplifies motion processing if their goal is to move towards a desired image and discuss dorsal and ventral stream processing of moving images in that context. Although many questions about view-based approaches to scene representation remain unanswered, the solutions are likely to be highly relevant to understanding biological 3-D vision.

Judgments of path, not heading, guide locomotion

To steer a course through the world, people are almost entirely dependent on visual information, of which a key component is optic flow. In many models of locomotion, heading is described as the fundamental control variable; however, it has also been shown that fixating points along or near one's future path could be the basis of an efficient control solution. Here, the authors aim to establish how well observers can pinpoint instantaneous heading and path, by measuring their accuracy when looking at these features while traveling along straight and curved paths. The results showed that observers could identify both heading and path accurately (similar to 3 degrees) when traveling along straight paths, but on curved paths they were more accurate at identifying a point on their future path (similar to 5 degrees) than indicating their instantaneous heading (similar to 13 degrees). Furthermore, whereas participants could track changes in the tightness of their path, they were unable to accurately track the rate of change of heading. In light of these results, the authors suggest it is unlikely that heading is primarily used by the visual system to support active steering.

Humans ignore motion and stereo cues in favor of a fictional stable world

As we move through the world, our eyes acquire a sequence of images. The information from this sequence is sufficient to determine the structure of a three-dimensional scene, up to a scale factor determined by the distance that the eyes have moved [1, 2]. Previous evidence shows that the human visual system accounts for the distance the observer has walked [3,4] and the separation of the eyes [5-8] when judging the scale, shape, and distance of objects. However, in an immersive virtual-reality environment, observers failed to notice when a scene expanded or contracted, despite having consistent information about scale from both distance walked and binocular vision. This failure led to large errors in judging the size of objects. The pattern of errors cannot be explained by assuming a visual reconstruction of the scene with an incorrect estimate of interocular separation or distance walked. Instead, it is consistent with a Bayesian model of cue integration in which the efficacy of motion and disparity cues is greater at near viewing distances. Our results imply that observers are more willing to adjust their estimate of interocular separation or distance walked than to accept that the scene has changed in size.

An automated calibration method for non-see-through head mounted displays

Accurate calibration of a head mounted display (HMD) is essential both for research on the visual system and for realistic interaction with virtual objects. Yet, existing calibration methods are time consuming and depend on human judgements, making them error prone. The methods are also limited to optical see-through HMDs. Building on our existing HMD calibration method [1], we show here how it is possible to calibrate a non-see-through HMD. A camera is placed inside an HMD displaying an image of a regular grid, which is captured by the camera. The HMD is then removed and the camera, which remains fixed in position, is used to capture images of a tracked calibration object in various positions. The locations of image features on the calibration object are then re-expressed in relation to the HMD grid. This allows established camera calibration techniques to be used to recover estimates of the display’s intrinsic parameters (width, height, focal length) and extrinsic parameters (optic centre and orientation of the principal ray). We calibrated a HMD in this manner in both see-through and in non-see-through modes and report the magnitude of the errors between real image features and reprojected features. Our calibration method produces low reprojection errors and involves no error-prone human measurements.

An automated calibration method for non-see-through head mounted displays

Accurate calibration of a head mounted display (HMD) is essential both for research on the visual system and for realistic interaction with virtual objects. Yet, existing calibration methods are time consuming and depend on human judgements, making them error prone, and are often limited to optical see-through HMDs. Building on our existing approach to HMD calibration Gilson et al. (2008), we show here how it is possible to calibrate a non-see-through HMD. A camera is placed inside a HMD displaying an image of a regular grid, which is captured by the camera. The HMD is then removed and the camera, which remains fixed in position, is used to capture images of a tracked calibration object in multiple positions. The centroids of the markers on the calibration object are recovered and their locations re-expressed in relation to the HMD grid. This allows established camera calibration techniques to be used to recover estimates of the HMD display's intrinsic parameters (width, height, focal length) and extrinsic parameters (optic centre and orientation of the principal ray). We calibrated a HMD in this manner and report the magnitude of the errors between real image features and reprojected features. Our calibration method produces low reprojection errors without the need for error-prone human judgements.