Effects of adaptation to apparent movement on recovery of structure from motion.

Past work on the recovery of three-dimensional structure from dynamic two-dimensional images has led to inconsistent conclusions regarding the contributions of the short-range and long-range motion processes. In the present experiments, subjects adapted to displays (either four lines or 50 randomly positioned pixels) whose spatiotemporal parameters were chosen to favor either the short-range or long-range process. Adaptation periods were followed by test displays that simulated the rotation of a four-pixel random object about the vertical gamma-axis. The dependent measure was the angle of rotation between successive frames of the rotation display at which percepts of three-dimensional structure broke down. Both the original data and derived measures based on best-fitting polynomials showed small but consistent effects: Compared to control conditions, adaptation to short-range motion reduced the angle at which percepts of structure broke down; adaptation to long-range motion increased them. It is suggested that both low-level (i.e. short-range) and high-level (long-range) processes contribute to the recovery of structure from motion.     

Retinal adaptation to object motion

Due to fixational eye movements, the image on the retina is always in motion, even when one views a stationary scene. When an object moves within the scene, the corresponding patch of retina experiences a different motion trajectory than the surrounding region. Certain retinal ganglion cells respond selectively to this condition, when the motion in the cell's receptive field center is different from that in the surround. Here we show that this response is strongest at the very onset of differential motion, followed by gradual adaptation with a time course of several seconds. Different subregions of a ganglion cell's receptive field can adapt independently. The circuitry responsible for differential motion adaptation lies in the inner retina. Several candidate mechanisms were tested, and the adaptation most likely results from synaptic depression at the synapse from bipolar to ganglion cell. Similar circuit mechanisms may act more generally to emphasize novel features of a visual stimulus.     

Spatiotopic perceptual maps in humans: evidence from motion adaptation

How our perceptual experience of the world remains stable and continuous despite the frequent repositioning eye movements remains very much a mystery. One possibility is that our brain actively constructs a spatiotopic representation of the world, which is anchored in external--or at least head-centred--coordinates. In this study, we show that the positional motion aftereffect (the change in apparent position after adaptation to motion) is spatially selective in external rather than retinal coordinates, whereas the classic motion aftereffect (the illusion of motion after prolonged inspection of a moving source) is selective in retinotopic coordinates. The results provide clear evidence for a spatiotopic map in humans: one which can be influenced by image motion.     

The interpretation of structure from motion.

The interpretation of structure from motion is examined from a computional point of view. The question addressed is how the three dimensional structure and motion of objects can be inferred from the two dimensional transformations of their projected images when no three dimensional information is conveyed by the individual projections. The following scheme is proposed: (i) divide the image into groups of four elements each; (ii) test each group for a rigid interpretation; (iii) combine the results obtained in (ii). It is shown that this scheme will correctly decompose scenes containing arbitrary rigid objects in motion, recovering their three dimensional structure and motion. The analysis is based primarily on the structure from motion theorem which states that the structure of four non-coplanar points is recoverable from three orthographic projections. The interpretation scheme is extended to cover perspective projections, and its psychological relevance is discussed.     

When motion is not perceived: evidence from adaptation and dynamical stability

Adaptation was used to probe the perceiver's activation state when either motion or nonmotion percepts are formed for bistable, single-element apparent motion stimuli. Although adaptation was not observed in every instance, when it was observed its effect was to increase the probability of both motion-to-nonmotion and nonmotion-to-motion switches, the time scale of adaptation corresponding to neurophysiological observations for directionally selective cortical cells (Giaschi et al. 1993). This susceptibility to de-stabilizing adaptation effects indicated that the nonmotion percept was not the result of inadequate stimulation producing subthreshold levels of motion detector activation; if that were the case, activation-dependent adaptation would have decreased the nonmotion-to-motion switching rate by reducing activation further below threshold. Above-threshold activation levels are therefore associated with both nonmotion and motion perceptual states, and the failure to perceive motion despite the presence of adequate motion detector stimulation can be attributed to inhibitory competition between detectors activated by motion-specifying stimulus information and detectors activated to similar levels by motion-independent stimulus information, consistent with the dynamical quality of single-element apparent motion.     

Contrast gain reduction in fly motion adaptation

In many species, including humans, exposure to high image velocities induces motion adaptation, but the neural mechanisms are unclear. We have isolated two mechanisms that act on directionally selective motion-sensitive neurons in the fly's visual system. Both are driven strongly by movement and weakly, if at all, by flicker. The first mechanism, a subtractive process, is directional and is only activated by stimuli that excite the neuron. The second, a reduction in contrast gain, is strongly recruited by motion in any direction, even if the adapting stimulus does not excite the cell. These mechanisms are well designed to operate effectively within the context of motion coding. They can prevent saturation at susceptible nonlinear stages in processing, cope with rapid changes in direction, and preserve fine structure within receptive fields.     

Attention, adaptation, and the motion aftereffect

Activation of the human visual motion area V5/MT was previously thought to be the basis of the motion aftereffect. New findings suggest that previous observations were confounded by attention and arousal, providing evidence that adaptation of directionally selective neurons in area V5/MT represents the fundamental substrate for the motion aftereffect.     

A model of motion adaptation and motion after-effects based upon principal component regression

A computational model to help explain effects of adaptation to moving signals is compared with established energy (linear regression) models of motion detection. The proposed model assumes that processed image signals are subject to error in both dimensions of space and time. This assumption constrains models of motion perception to be based upon principal component regression rather than linear regression. It is shown that response suppression of model complex cell neurons that input into the model may account for (1) increases in perceived speed after adaptation to static patterns and testing with slowly moving patterns, (2) significant increases in perceived speed after adaptation to patterns moving at a medium speed and testing at high speed, and (3) decreases in perceived speed in the opponent direction to a quickly moving adapting signal. Neither of predictions (2) or (3) are general features of established accounts of motion detection by visual processes based upon linear regression. Comparisons of the proposed model's speed transfer function with existing psychophysical data suggests that the visual system processes motion signals with the tacit assumption that image measurements are subject to error in both space and time. Received: 24 January 2000 / Accepted in revised form: 8 May 2000     

Predicting enzyme class from protein structure without alignments

Methods for predicting protein function from structure are becoming more important as the rate at which structures are solved increases more rapidly than experimental knowledge. As a result, protein structures now frequently lack functional annotations. The majority of methods for predicting protein function are reliant upon identifying a similar protein and transferring its annotations to the query protein. This method fails when a similar protein cannot be identified, or when any similar proteins identified also lack reliable annotations. Here, we describe a method that can assign function from structure without the use of algorithms reliant upon alignments. Using simple attributes that can be calculated from any crystal structure, such as secondary structure content, amino acid propensities, surface properties and ligands, we describe each enzyme in a non-redundant set. The set is split according to Enzyme Classification (EC) number. We combine the predictions of one-class versus one-class support vector machine models to make overall assignments of EC number to an accuracy of 35% with the top-ranked prediction, rising to 60% accuracy with the top two ranks. In doing so we demonstrate the utility of simple structural attributes in protein function prediction and shed light on the link between structure and function. We apply our methods to predict the function of every currently unclassified protein in the Protein Data Bank.     

Two-dimensional motion perception without feature tracking

Feature-tracking explanations of 2D motion perception are fundamentally distinct from motion-energy, correlation, and gradient explanations, all of which can be implemented by applying spatiotemporal filters to raw image data. Filter-based explanations usually suffer from the aperture problem, but 2D motion predictions for moving plaids have been derived from the intersection of constraints (IOC) imposed by the outputs of such filters, and from the vector sum of signals generated by such filters. In most previous experiments, feature-tracking and IOC predictions are indistinguishable. By constructing plaids in apparent motion from missing-fundamental gratings, we set feature-tracking predictions in opposition to both IOC and vector-sum predictions. The perceived directions that result are inconsistent with feature tracking. Furthermore, we show that increasing size and spatial frequency in Type 2 missing-fundamental plaids drives perceived direction from vector-sum toward IOC directions. This reproduces results that have been used to support feature-tracking, but under experimental conditions that rule it out. We discuss our data in the context of a Bayesian model with a gradient-based likelihood and a prior favoring slow speeds. We conclude that filter-based explanations alone can explain both veridical and non-veridical 2D motion perception in such stimuli.     

The computation of structure from fixed-axis motion: rigid structures

We show that three distinct orthographic views of three points in a rigid configuration are compatibel with at most 64 interpretations of the three-dimensional structure and motion of the points. If, in addition, one assumes that the three points spin about a fixed axis over the three views, then with probability one there is a unique three-dimensional interpretation (plus a reflection). Moreover the probability of false targets is zero. In the special case that the axis of rotation is parallel to the image plane three views of the three points are sufficient to obtain at most two interpretations (plus reflections)-unless one assumes the angular velocity about the axis is constant, in which case three views of two points are sufficient to determine a unique interpretation. Closed form solutions are obtained for each of these cases. The systems of equations studied here are in each case overconstraining (i.e. there are more independent equations than unknowns) and are amenable to solution by nonlinear programming. These two properties make possible the construction of noise insensitive algorithms for computer vision systems. Our uniqueness proofs employ the Principle of upper semicontinuity, a principle which underlies a general mathematical framework for the analysis of solutions to overconstraining systems of equations.     

The computation of structure from fixed-axis motion: Nonrigid structures

We show that an assumption of rigidity or quasi-rigidity is not necessary, in principle, for the computation of three-dimensional structure and motion from changing retinal images. In particular, we show that the three-dimensional structure of certain nonrigid objects, namely objects whose texture elements rotate about a common axis but at varying angular velocities, can in principle be computed from three successive retinal images of four texture elements, or from four successive images of two texture elements. We then show that in both cases the computed structure matches the actual structure of the object with probability one.     

Complex adaptation and system structure

The structural organization of biological systems is one of nature’s most fascinating aspects, but its origin and functional role is not yet fully understood. For instance, basic adaptational mechanisms like genetic mutation and Hebbian adaptation seem to be generic and invariant across many species and are, on their own, fairly well investigated and understood. However, it is the organism’s structure – the representations these mechanisms act upon – that bears the complex functional effects of these mechanisms. While typical technical approaches to system design require detailed problem models and suffer from the need to explicitly take care of all possible cases, the organization of biological systems seems to induce inherent adaptability, flexibility and robustness. In this discussion paper we address the concept of structured variability, particularly the role of system structure as implementing a certain representation on which basic variational mechanisms act on. The functional adaptability (or search distribution) depends crucially on this representation.     

Seeing blur: 'motion sharpening' without motion

It is widely supposed that things tend to look blurred when they are moving fast. Previous work has shown that this is true for sharp edges but, paradoxically, blurred edges look sharper when they are moving than when stationary. This is 'motion sharpening'. We show that blurred edges also look up to 50% sharper when they are presented briefly (8-24 ms) than at longer durations (100-500 ms) without motion. This argues strongly against high-level models of sharpening based specifically on compensation for motion blur. It also argues against a recent, low-level, linear filter model that requires motion to produce sharpening. No linear filter model can explain our finding that sharpening was similar for sinusoidal and non-sinusoidal gratings, since linear filters can never distort sine waves. We also conclude that the idea of a 'default' assumption of sharpness is not supported by experimental evidence. A possible source of sharpening is a nonlinearity in the contrast response of early visual mechanisms to fast or transient temporal changes, perhaps based on the magnocellular (M-cell) pathway. Our finding that sharpening is not diminished at low contrast sets strong constraints on the nature of the nonlinearity.     

Isotropic integration of binocular disparity and relative motion in the perception of three-dimensional shape

Richards (1985) showed that veridical three-dimensional shape may be recovered from the integration of binocular disparity and retinal motion information, but proposed that this integration may only occur for horizontal retinal motion. Psychophysical evidence supporting the combination of stereo and motion information is limited to the case of horizontal motion (Johnston et al., 1994), and has been criticised on the grounds of potential object boundary cues to shape present in the stimuli. We investigated whether veridical shape can be recovered under more general conditions. Observers viewed cylinders that were defined by binocular disparity, two-frame motion or a combination of disparity and motion, presented at simulated distances of 30 cm, 90 cm or 150 cm. Horizontally and vertically oriented cylinders were rotated about vertical and horizontal axes. When rotation was about the cylinder's own axis, no boundary cues to shape were introduced. Settings were biased for the disparity and two-frame motion stimuli, while more veridical shape judgements were made under all conditions for combined cue stimuli. These results demonstrate that the improved perception of three-dimensional shape in these stimuli is not a consequence of the presence of object boundary cues, and that the combination of disparity and motion is not restricted to horizontal image motion.     

Comparison of various motion stimuli on motion sickness and acquisition of adaptation in Suncus Murinus

Effects of various types of motion stimuli were compared to investigate optimum method to elicit motion sickness and adaptation in Suncus murinus (suncus). Three different direction of shaking in the horizontal plane, back and forth, right and left and revolving, induced emetic response to the similar extent. However, vertical shaking was far less effective in inducing motion sickness. Mild and severe horizontal shaking (15 min per day) was continued for 14 days and emetic response to standard motion stimulus was compared before and after the training. The severe daily acceleration strongly depressed the susceptibility to motion stimulus. The mild acceleration which was not emetic stimulus in itself also remarkably attenuated the vomiting response to standard motion stimulus. These results indicate that 1) the emetic responsiveness of the suncus does not depend on the modes of shaking as long as the direction is in the horizontal plane, 2) the suncus is relatively refractory to the vertical linear acceleration and 3) the adaptation to motion stimulus does not develop on the latest peripheral steps of the vomiting reflex pathways.     

Visual motion induces a forward prediction of spatial pattern

Cortical motion analysis continuously encodes image velocity but might also be used to predict future patterns of sensory input along the motion path. We asked whether this predictive aspect of motion is exploited by the human visual system. Targets can be more easily detected at the leading as compared to the trailing edge of motion [1], but this effect has been attributed to a nonspecific boost in contrast gain at the leading edge, linked to motion-induced shifts in spatial position [1-4]. Here we show that the detectability of a local sinusoidal target presented at the ends of a region containing motion is phase dependent at the leading edge, but not at the trailing edge. These two observations rule out a simple gain control mechanism that modulates contrast energy and passive filtering explanations, respectively. By manipulating the relative orientation of the moving pattern and target, we demonstrate that the resulting spatial variation in detection threshold along the edge closely resembles the superposition of sensory input and an internally generated predicted signal. These findings show that motion induces a forward prediction of spatial pattern that combines with the cortical representation of the future stimulus.     

Mapping coral reefs using consumer-grade drones and structure from motion photogrammetry techniques

We propose a novel technique to measure the small-scale three-dimensional features of a shallow-water coral reef using a small drone equipped with a consumer-grade camera, a handheld GPS and structure from motion (SfM) algorithms. We used a GoPro HERO4 with a modified lens mounted on a DJI Phantom 2 drone (maximum total take-off weight <2 kg) to perform a 10 min flight and collect 306 aerial images with an overlap equal or greater than 90%. We mapped an area of 8380 m², obtaining as output an ortho-rectified aerial photomosaic and a bathymetric digital elevation model (DEM) with a resolution of 0.78 and 1.56 cm pixel⁻¹, respectively. Through comparison with airborne LiDAR data for the same area, we verified that the location of the ortho-rectified aerial photomosaic is accurate within ~1.4 m. The bathymetric difference between our DEM and the LiDAR dataset is −0.016 ± 0.45 m (1σ). Our results show that it is possible, in conditions of calm waters, low winds and minimal sun glint, to deploy consumer-grade drones as a relatively low-cost and rapid survey technique to produce multispectral and bathymetric data on shallow-water coral reefs. We discuss the utility of such data to monitor temporal changes in topographic complexity of reefs and associated biological processes.