ArticlePDF Available

Abstract

We studied whether the blur/sharpness of an occlusion boundary between a sharply focused surface and a blurred surface is used as a relative depth cue. Observers judged relative depth in pairs of images that differed only in the blurriness of the common boundary between two adjoining texture regions, one blurred and one sharply focused. Two experiments were conducted; in both, observers consistently used the blur of the boundary as a cue to relative depth. However, the strength of the cue, relative to other cues, varied across observers. The occlusion edge blur cue can resolve the near/far ambiguity inherent in depth-from-focus computations.
... In the current study, we chose to focus on blur as a depth cue. Blur has been shown to affect depth perception, as it has been found to be correlated with the amount of surface blur (e.g., Marshall et al., 1996;Mather & Smith, 2002;O'Shea et al., 1997). Notably, blur does not provide the sign for the distance of the object from the participant (i.e., whether it is closer or farther away), as it occurs in normal vision when the object is near or far from the plane of focus. ...
... Notably, blur does not provide the sign for the distance of the object from the participant (i.e., whether it is closer or farther away), as it occurs in normal vision when the object is near or far from the plane of focus. It has been suggested that resolving the sign problem with blurry surfaces occurs through occlusion edge blur, as the blurriness of the border between two textures is affiliated with the texture with a similar amount of blur, leading to the perception of occlusion where that texture is perceived to be in front of the other one (Marshall et al., 1996;Mather, 1996). Later, Palmer and Brooks (2008) reported a related grouping effect (edge-region grouping [ERG]), extending it to surfaces comprised of discrete elements and a line that functioned as a border due to similarity by blur. ...
... Importantly, sharp elements were perceived as closer the majority of the time, irrelevant of whether they formed the background or the figure, with the exception of the blurry lines in Experiment 3b. Previous research has suggested that the resolution of the sign problem in depth from blur is achieved by the grouping of a border with a surface according to their similarity in blur (e.g., Marshall et al., 1996;Mather, 1996;Palmer & Brooks, 2008). However, we provide here new evidence for the efficiency of blur as a depth cue in grouping organizations that do not include borders between surfaces. ...
Article
Previous studies provided evidence in support of attention operating in three-dimensional space, and the iterative and multistage nature of organizational processes in relation to attention and depth. We investigated depth perception and attentional demands in grouping organizations that contain blur as a depth cue. Contrary to previous studies, in our displays, no depth from occlusion could be implied from a shared border between groups or surfaces. To evaluate depth perception, subjective reports were collected where participants indicated which elements, blurry or sharp, they perceived as closer. To examine whether depth perception from blur can alleviate attentional demands, we used an inattention paradigm. We presented displays of grouping organizations by collinearity or color similarity that were previously found to require attention and added blur to the figure or the background elements to generate depth perception. In addition, we presented similar displays containing grouping by blur similarity as a single cue. We hypothesized that adding blur would facilitate the segmentation of element groups due to their perceived depth, which might lead to a diminished demand for attention. Our results confirmed that blur led to depth perception, and that sharp elements were perceived as closer more frequently than blurry elements. Thus, these results provide novel evidence for depth from blur in grouping where no inference of occlusion can be derived from a border. However, although the results suggest that blur information was processed under inattention, little evidence was found for decreased attentional demands for grouping processes in the presence of blur.
... Depending on the number, category, orientation, and position of objects, the boundaries of objects are elusive; no simple priors can be applied to recover the foreground and background in the scene. Occlusion relationship reasoning of objects from monocular images reveals relative depth differences among the objects in the scene, which is fundamental in computer vision applications, such as mobile robots [17,27,2], object detection [8,4,31,20,5], segmentation [8,37,6,30,13,38,21], monocular depth * corresponding author. estimation/ordering [22,24,23,25,16,36], and 3D reconstruction [29,19,9]. ...
Preprint
Full-text available
Retrieving occlusion relation among objects in a single image is challenging due to sparsity of boundaries in image. We observe two key issues in existing works: firstly, lack of an architecture which can exploit the limited amount of coupling in the decoder stage between the two subtasks, namely occlusion boundary extraction and occlusion orientation prediction, and secondly, improper representation of occlusion orientation. In this paper, we propose a novel architecture called Occlusion-shared and Path-separated Network (OPNet), which solves the first issue by exploiting rich occlusion cues in shared high-level features and structured spatial information in task-specific low-level features. We then design a simple but effective orthogonal occlusion representation (OOR) to tackle the second issue. Our method surpasses the state-of-the-art methods by 6.1%/8.3% Boundary-AP and 6.5%/10% Orientation-AP on standard PIOD/BSDS ownership datasets. Code is available at https://github.com/fengpanhe/MT-ORL.
... They showed that increasing the window size reduces the depth resolution, due to the over-smoothing effects. On the other hand, small windows increase the image's sensitivity to noise [8,9]. Pertuz et al. [7] suggested that the optimum window size is a trade-off between spatial resolution and robustness. ...
Article
Full-text available
This paper reports the precision of shape-from-focus (SFF) imaging according to the texture frequencies and window sizes of a focus measure. SFF is one of various depth measurement techniques for optical imaging, such as microscopy and endoscopy. SFF measures the depth of an object according to focus measure, which is generally computed with a fixed window. The window size affects the performance of SFF and should be adjusted for the texture of an object. In this study, we investigated the precision difference of SFF in texture frequencies and by window size. Two experiments were performed: precision validation in texture frequencies with a fixed window size, and precision validation in various window sizes related to pixel-cycle lengths. The first experimental results showed that a smaller window size could not provide a correct focus measure, and the second results showed that a window size that is approximately equal to a pixel-cycle length of the texture could provide better precision. These findings could potentially contribute to determining the appropriate window size of focus measure operation in shape-from-focus reconstruction.
... These images often have unavoidably different low-level image features, and these low-level image features themselves could affect the perception of distance and depth. These features include size [32][33][34], spatial-frequency [35][36][37], contrast [38][39][40], color [41][42][43], and configuration [44,45]. These effects of low-level image features can be confounded with familiarity in measurements of distance and depth. ...
Article
Full-text available
Bishop Berkeley suggested that the distance of an object can be estimated if the object’s size is familiar to the observer. It has been suggested that humans can perceive the distance of the object by using such “familiarity” information, but most or many of the prior experiments that found an effect of familiarity were not designed to minimize or eliminate potential influences of: higher cognitive factors on the observers’ responses, or the influences of low-level image features in the visual stimuli used. We looked for the familiarity effect in two experiments conducted both in Russia and Japan. The visual stimuli used were images of three coins used in Russia and Japan. The participants’ depth perception was measured with a multiple-choice task testing the perceived depth-order of the coins. Our expectation was that any effect of “familiarity” on depth perception would only be observed with the coins of the participant’s country. We expected a substantial familiarity effect based on our meta-analysis of the “familiarity” effects observed in prior experiments. But, our results in both experiments showed that the familiarity effect was virtually zero. These findings suggest that the importance of a familiarity effect in depth perception should be reconsidered.
... Evidence for the use of image blur in depth perception has been reported by Mather [9] and by Marshall et al [4] . Their papers described experiments on ambiguous figure-ground stimuli, containing two regions of texture separated by a wavy boundary. ...
Chapter
Full-text available
In recent years, 3D model visualization techniques have made enormous progress. This evolution has not only touched the technical side but also the hardware side. It is no longer necessary to have expensive machines to see a world in 3D; a simple computer can do the trick. Historically, research has focused on the development of 3D information and acquisition techniques from scenes and objects. These acquisition methods require expertise and complex calibration procedures whenever the acquisition system was used. All this creates an important demand for flexibility in these methods of acquisition because of these different factors, many techniques have emerged. Many of them only need a camera and a computer to create a 3D world from a scene.
Thesis
Full-text available
Absolutely measuring angle of rotation and measuring torque are of great importance in industry. So far, the acquisition of both variables is done separately. Within the scope of this thesis, a non-contact measurement method for the determination of the absolute rotation angle and torque is investigated. In order to detect the torque of a mechanical shaft contactless, two coded markings in the circumferential direction and corresponding distance are applied with a laser, which are red by means of an optical system. Both markings encode the absolute angle of rotation within 360°, so that when the two angles of rotation are detected, an angle difference and thus torque are inferred. Further development of existing codings as well as the detectability of the absolute rotation angles and the torque calculated therefrom are described in this work. In addition, the detection of the circularity and the effect of this on the angle measurement in the context of the scientific question is examined. For this purpose, both the model presentation and experimental measurement results are described in the work. The simultaneous detection of the circularity offers the possibility of determining the angle-dependent change in distance, so that an application in condition-based maintenance (e.g. of bearings) is conceivable. Furthermore, preliminary considerations for industrial applicability are included in order to enable a compact design and further optimize data processing. In experiments with an engine test bench, the implemented prototypical measuring system is investigated and discussed in industrial practice. At a speed of 500 1/min, the general feasibility of optical torque measurement was demonstrated with an error of up to 11 %.
Article
Immersive computer graphics systems strive to generate perceptually realistic user experiences. Current-generation virtual reality (VR) displays are successful in accurately rendering many perceptually important effects, including perspective, disparity, motion parallax, and other depth cues. In this article, we introduce ocular parallax rendering, a technology that accurately renders small amounts of gaze-contingent parallax capable of improving depth perception and realism in VR. Ocular parallax describes the small amounts of depth-dependent image shifts on the retina that are created as the eye rotates. The effect occurs because the centers of rotation and projection of the eye are not the same. We study the perceptual implications of ocular parallax rendering by designing and conducting a series of user experiments. Specifically, we estimate perceptual detection and discrimination thresholds for this effect and demonstrate that it is clearly visible in most VR applications. Additionally, we show that ocular parallax rendering provides an effective ordinal depth cue and it improves the impression of realistic depth in VR.
Article
Full-text available
The injection rate of particles at the front of interplanetary shocks has usually been studied locally, around the shock passage by the observer's position, but little is known of how the efficiency of the particle-acceleration process evolves as the shock propagates from the Sun to the Earth. In many events accelerated particles are observed long in advance of the arrival of the shock (from 5 hours to 2 days), and they show large anisotropies. We have used a compound shock-particle model to derive the injection rate of particles at the shock front and their energy spectrum, as a function of time, by fitting the observed particle fluxes and anisotropies between 100 and 1000 keV. We have studied three individual low-energy particle events taken as representatives of West, Central Meridian and East events.
Article
Full-text available
We report a spatial frequency influence on perceptual organization. If regions filled with relatively higher spatial frequency sinusoidal gratings are adjacent to regions containing relatively lower spatial frequency gratings, the regions with the higher frequency will appear closer in depth than those containing the lower frequency. This depth effect is evident even when conflicting stereo information is present.
Conference Paper
One of the major unsolved problems in designing an autonomous agent [robot] that must function in a complex, moving environment is obtaining reliable, real-time depth information, preferably without the limitations of active scanners. Stereo remains computationally intensive and prone to severe errors, the use of motion information is still quite experimental, and autofocus schemes can measure depth at only one point at a time. We examine a novel source of depth information: focal gradients resulting from the limited depth of field inherent in most optical systems. We prove that this source of information can be used to make reliable depth maps of useful accuracy with relatively minimal computation. Experiments with realistic imagery show that measurement of these optical gradients can potentially provide depth information roughly comparable to stereo disparity or motion parallax, while avoiding image-to-image matching problems. A potentially real-time version of this algorithm is described.
Article
This paper examines a novel source of depth information: focal gradients resulting from the limited depth of field inherent in most optical systems. Previously, autofocus schemes have used depth of field to measured depth by searching for the lens setting that gives the best focus, repeating this search separately for each image point. This search is unnecessary, for there is a smooth gradient of focus as a function of depth. By measuring the amount of defocus, therefore, we can estimate depth simultaneously at all points, using only one or two images. It is proved that this source of information can be used to make reliable depth maps of useful accuracy with relatively minimal computation. Experiments with realistic imagery show that measurement of these optical gradients can provide depth information roughly comparable to stereo disparity or motion parallax, while avoiding image-to-image matching problems.
Article
A new method for extracting depth information from 2-dimensional images is propsed. The depth structure of a scene is computed using measurements of the degree of blur in an image which is only in parts sharply focused. Preliminary tests suggest that the method is surprisingly powerful and may have interesting applications.
Article
In recent times a great deal of interest has been shown, amongst the computer vision and robotics research community, in the acquisition of range data for supporting scene analysis leading to remote (noncontact) determination of configurations and space filling extents of three-dimensional object assemblages. This paper surveys a variety of approaches to generalized range finding and presents a perspective on their applicability and shortcomings in the context of computer vision studies.
Article
Image regions corresponding to partially hidden objects are enclosed by two types of bounding contour: those inherent to the object itself (intrinsic) and those defined by occlusion (extrinsic). Intrinsic contours provide useful information regarding object shape, whereas extrinsic contours vary arbitrarily depending on accidental spatial relationships in scenes. Because extrinsic contours can only degrade the process of surface description and object recognition, it is argued that they must be removed prior to a stage of template matching. This implies that the two types of contour must be distinguished relatively early in visual processing and we hypothesize that the encoding of depth is critical for this task. The common border is attached to and regarded as intrinsic to the closer region, and detached from and regarded as extrinsic to the farther region. We also suggest that intrinsic borders aid in the segmentation of image regions and thus prevent grouping, whereas extrinsic borders provide a linkage to other extrinsic borders and facilitate grouping. Support for these views is found in a series of demonstrations, and also in an experiment where the expected superiority of recognition was found when partially sampled faces were seen in a back rather than a front stereoscopic depth plane.
Article
In three experiments we examined the influence of spatial frequency on perceptual organization. In each experiment a pattern was tested that was ambiguous in terms of figure and ground. In each experiment, the stimuli were 20 variations of the pattern, which were generated by filling the two regions of the pattern with horizontal sine wave gratings differing in spatial frequency. Five spatial frequencies were tested: 0.5, 1, 2, 4, and 8 cycles per degree. The response measure was the percentage of response time one of the regions was seen as the figure. This region was seen as the figure a higher percentage of the time in those stimuli where it contained the relatively higher spatial frequency sine wave grating compared with those stimuli where it contained the relatively lower spatial frequency sine wave grating.