Unexpected Objects Research Articles

When the WRENCH turns a few heads: Expectation and semantic relatedness in inattentional blindness

Semantic relatedness and expectation were investigated in inattentional blindness—failure to perceive an unexpected object in plain sight when attention is engaged elsewhere. Participants named primary-task pictures and ignored distractor pictures. Four trials preceded a ‘critical’ trial where an unexpected six-letter-word appeared at fixation, simultaneously with the pictures. In Experiment 1, we found robust effects for both in-lab and on-line-Zoom methodology. More participants reported the unexpected word semantically-related to the primary-task pictures than a semantically-unrelated word. In Experiment 2, expectations were violated, by changing the semantic category of the primary-task pictures. More participants reported the unexpected word semantically-related to the unexpected picture category than a semantically-unrelated word. When attentional resources are consumed by a task, a violation to task expectations is not enough to reorient attention to an unexpected word. Attention reorients to what is meaningful to the task, and what is meaningful is updated in light of unexpected information.

A search for transient, monochromatic light in a 6-deg swath along the galactic plane

ABSTRACT We searched the Milky Way Plane along a 6-deg swath for pulses of monochromatic light as faint as 15th mag (V band) using a wide-field telescope equipped with a prism. Pulses with duration less than 1 s that occur more often than once every 10 min would be detected, and pulses arriving less frequently would be detected with proportionally lower probability. A ‘difference-image’ algorithm revealed 36 monochromatic sources. Subsequent assessment showed all were simply astrophysical objects emitting known atomic emission lines. No unexplainable monochromatic emission, pulsed or continuous, was detected. The detection threshold corresponds to an ∼70 GW laser having a diffraction-limited 10-m aperture 1 kpc away (depending on wavelength). Past all-sky optical and radio-wave surveys revealed thousands of unexpected objects exhibiting extraordinary spectral emission, but none were technological. Hypotheses of our Milky Way Galaxy teeming with advanced life must be demoted.

World: Resisting singular objectification in pursuit of multiplicity

This paper contrasts a singular construction of the ‘world’ as an object, often presented to students through representational objects (such as atlases, maps, globes), with a more expansive orientation toward the idea of ‘world’ and how it might enter the classroom. By engaging with Hannah Arendt’s concept of natality, I consider a view of the world as encompassing the many unexpected objects, ideas, relationships, and experiences students might bring forth: how might educators make space for the many ‘worlds’ students might bring into the classroom with them? For what reasons might it be important to see world construction as an active process instead of a prescribed one? What new opportunities for worlding are made possible when the ‘world’ is seen as many, potentially disruptive, things, instead of as a singular ‘object’? Following these questions, in this paper ‘world’ is presented as multiple: something we attempt to reduce to the singular; something plural - the objects that punctuate our lives in unexpected ways; something we build alone and together; a speculative encounter with others in complex relation; and, ultimately, an act of love.

Unexpected object detection based on class correlation in semantic segmentation for automatic driving scenes

The detection of out-of-distribution (OoD) samples in semantic segmentation is crucial for autonomous driving, as deep learning models are typically trained under the assumption of a closed environment, whereas the real world presents an open and diverse set of scenarios. Existing methods employ uncertainty estimation, image reconstruction, and other techniques for OoD sample detection. We have observed that different classes may exhibit connections and associations in varying contexts. For example, objects encountered by autonomous vehicles differ in rural road scenes compared to urban environments, and the likelihood of encountering novel objects varies. This aspect is missing in current anomaly detection methods and is vital for OoD sample detection. Existing approaches solely consider the relative significance of each prediction class, overlooking the inter-object correlation. Although prediction scores (e.g., max logits) obtained from the segmentation network are applicable for OoD sample detection, the same problem persists, particularly for OoD objects. To address this issue, we propose the utilization of the Mahalanobis distance of max logits to evaluate the final predicted score. By calculating the Mahalanobis distance, the paper aims to uncover correlations between different classes, thus enhancing the effectiveness of OoD detection. To this end, we also extend the state-of-the-art segmentation model, DeepLabV3+, to enable OoD sample detection in this paper. Specifically, this paper proposes a novel backbone network, SOD-ResNet101, for extracting contextual and multi-scale semantic information, leveraging the class correlation feature of the Mahalanobis distance to enhance the detection performance of out-of-distribution objects. Notably, our approach eliminates the need for external datasets or separate network training, making it highly applicable to existing pretraining segmentation models.

Are Familiar Objects More Likely to Be Noticed in an Inattentional Blindness Task?

People often fail to notice the presence of unexpected objects when their attention is engaged elsewhere. In dichotic listening tasks, for example, people often fail to notice unexpected content in the ignored speech stream even though they occasionally do notice highly familiar stimuli like their own name (the "cocktail party" effect). Some of the first studies of inattentional blindness were designed as a visual analog of such dichotic listening studies, but relatively few inattentional blindness studies have examined how familiarity affects noticing. We conducted four preregistered inattentional blindness experiments (total N = 1700) to examine whether people are more likely to notice a familiar unexpected object than an unfamiliar one. Experiment 1 replicated evidence for greater noticing of upright schematic faces than inverted or scrambled ones. Experiments 2-4 tested whether participants from different pairs of countries would be more likely to notice their own nation's flag or petrol company logo than those of another country. These experiments repeatedly found little or no evidence that familiarity affects noticing rates for unexpected objects. Frequently encountered and highly familiar stimuli do not appear to overcome inattentional blindness.

Individual differences in inattentional blindness.

People often fail to notice unexpected objects and events when they are performing an attention-demanding task, a phenomenon known as inattentional blindness. We might expect individual differences in cognitive ability or personality to predict who will and will not notice unexpected objects given that people vary in their ability to perform attention-demanding tasks. We conducted a comprehensive literature search for empirical inattentional blindness reports and identified 38 records that included individual difference measures and met our inclusion criteria. From those, we extracted individual difference effect sizes for 31 records which included a total of 74 distinct, between-groups samples with at least one codable individual difference measure. We conducted separate meta-analyses of the relationship between noticing/missing an unexpected object and scores on each of the 14 cognitive and 19 personality measures in this dataset. We also aggregated across personality measures reflecting positive/negative affectivity or openness/absorption and cognitive measures of interference, attention breadth, and memory. Collectively, these meta-analyses provided little evidence that individual differences in ability or personality predict noticing of an unexpected object. A robustness analysis that excluded samples with extremely low numbers of people who noticed or missed produced similar results. For most measures, the number of samples and the total sample sizes were small, and larger studies are needed to examine individual differences in inattentional blindness more systematically. However, the results are consistent with the idea that noticing of unexpected objects or events differs from deliberate attentional control tasks in that it is not reliably predicted by individual differences in cognitive ability.

A Parallel Open-World Object Detection Framework with Uncertainty Mitigation for Campus Monitoring

The recent advancements in artificial intelligence have brought about significant changes in education. In the context of intelligent campus development, target detection technology plays a pivotal role in applications such as campus environment monitoring and the facilitation of classroom behavior surveillance. However, traditional object detection methods face challenges in open and dynamic campus scenarios where unexpected objects and behaviors arise. Open-World Object Detection (OWOD) addresses this issue by enabling detectors to gradually learn and recognize unknown objects. Nevertheless, existing OWOD methods introduce two major uncertainties that limit the detection performance: the unknown discovery uncertainty from the manual generation of pseudo-labels for unknown objects and the known discrimination uncertainty from perturbations that unknown training introduces to the known class features. In this paper, we introduce a Parallel OWOD Framework with Uncertainty Mitigation to alleviate the unknown discovery uncertainty and the known discrimination uncertainty within the OWOD task. To address the unknown discovery uncertainty, we propose an objectness-driven discovery module to focus on capturing the generalized objectness shared among various known classes, driving the framework to discover more potential objects that are distinct from the background, including unknown objects. To mitigate the discrimination uncertainty, we decouple the learning processes for known and unknown classes through a parallel structure to reduce the mutual influence at the feature level and design a collaborative open-world classifier to achieve high-performance collaborative detection of both known and unknown classes. Our framework provides educators with a powerful tool for effective campus monitoring and classroom management. Experimental results on standard benchmarks demonstrate the framework’s superior performance compared to state-of-the-art methods, showcasing its transformative potential in intelligent educational environments.

Similarity of an unexpected object to the attended and ignored objects affects noticing in a sustained inattentional blindness task.

When focusing attention on some objects and ignoring others, people often fail to notice the presence of an additional, unexpected object (inattentional blindness). In general, people are more likely to notice when the unexpected object is similar to the attended items and dissimilar from the ignored ones. Perhaps surprisingly, current evidence suggests that this similarity effect results almost entirely from dissimilarity to the ignored items, and it remains unclear whether similarity to the attended items affects noticing. Other aspects of similarity have not been examined at all, including whether the similarity of the attended and ignored items to each other affects noticing of a distinct unexpected object. We used a sustained inattentional blindness task to examine all three aspects of similarity. Experiment 1 (n = 813) found no evidence that increasing the similarity of the attended and ignored items to each other affected noticing of an unexpected object. Experiment 2 (n = 610) provided some of the first compelling evidence that similarity to the attended items - in addition to the ignored items - affects noticing. Experiment 3 (n = 1,044) replicated that pattern and showed that noticing rates varied with the degree of similarity to the ignored shapes but not to the attended shapes, suggesting that suppression of ignored items functions differently from the enhancement of attended items.

An investigation of inattentional blindness using gaze and frequency tagging.

Inattentional blindness (IB) occurs when a salient object presented in plain sight goes unnoticed when its appearance is unexpected. Across two experiments, participants completed a classic dynamic IB task while eye movements and steady-state visual evoked potential (SSVEP) responses were continually recorded. This allowed us to measure the modulation of gaze and brain-based indices of attention during IB. While an SSVEP response to all stimuli including the unexpected object was attained, only gaze measures were able to discriminate noticers from nonnoticers. Experiment 1 used a prototypical sustained IB task and found that gaze toward the unexpected object was largely unrelated to noticing that object. Experiment 2 manipulated the contrast of the target and distractor stimuli, and instead observed a tight concordance between gazing at the unexpected object and reporting its presence. This task-based variability in gaze deployment is consistent with the broader literature and cumulatively delineates the challenges faced in translating lab-based IB research from the bench to the bedside. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Using Feedforward to Reveal Interaction Possibilities in Virtual Reality

In virtual reality (VR), interactions may fail when users encounter new, unknown, or unexpected objects. We propose using feedforward in VR to help users interact with objects by revealing how such objects work. Feedforward lets users know what to do and how to do it by showing the available actions and outcomes before an interaction. In this article, we first chart the design space of feedforward in VR and illustrate how to design feedforward for specific VR interactions. We discuss starting the feedforward, previewing actions and outcomes, and returning the virtual world to its state before the feedforward. Second, we implement three real-world VR applications to show how feedforward can be applied to multistep interactions, perceived interactivity, and discoverability. Third, we conduct an evaluation of the design space with 14 VR experts to understand its usefulness. Finally, we summarize the findings of our work on VR feedforward in 15 guidelines.

Poster Session: Eye-tracking while learning greebles.

There has been a lot of attention on so-called 'adversarial images' that fool machine learning models (MLM). Manipulations to an image or including an unexpected object in a scene can severely disrupt object labeling and parsing by state-of-the-art MLM such as convolutional neural networks (CNN). In contrast, human observers may not even notice these changes, highlighting a significant gap between the robustness of human vision and CNNs. One well-studied class of novel objects are 'Greebles,' introduced by Gauthier and Tarr (1997; Gauthier et al., 1998). A large number of 'families', 'genders' and individuals can be created by systematically varying the arrangements and shapes of Greeble components. We report on eye-tracking (SMI iRed 250) participants as they are taught to recognize Greebles following the procedure described by Gauthier, Tarr and colleagues. Greeble expertise is assessed using a naming task and a verification task. Once an individual became a Greeble 'expert' we assessed their tolerance for rotations of the Greebles around the y-axis. The pattern of eye movements when learning and being tested on the Greebles are compared to the output of CNNs trained to recognize Greebles.

Aging attenuates the memory advantage for unexpected objects in real-world scenes

Across the adult lifespan memory processes are subject to pronounced changes. Prior knowledge and expectations might critically shape functional differences; however, corresponding findings have remained ambiguous so far. Here, we chose a tailored approach to scrutinize how schema (in-)congruencies affect older and younger adults’ memory for objects embedded in real-world scenes, a scenario close to everyday life memory demands. A sample of 23 older (52–81 years) and 23 younger adults (18–38 years) freely viewed 60 photographs of scenes in which target objects were included that were either congruent or incongruent with the given context information. After a delay, recognition performance for those objects was determined. In addition, recognized objects had to be matched to the scene context in which they were previously presented. While we found schema violations beneficial for object recognition across age groups, the advantage was significantly less pronounced in older adults. We moreover observed an age-related congruency bias for matching objects to their original scene context. Our findings support a critical role of predictive processes for age-related memory differences and indicate enhanced weighting of predictions with age. We suggest that recent predictive processing theories provide a particularly useful framework to elaborate on age-related functional vulnerabilities as well as stability.

Classification Model Based on U-Net for Crack Detection from Asphalt Pavement Images

The purpose of our study is to detect cracks accurately from asphalt pavement surface images, which includes unexpected objects, non-uniform illumination, and irregularities in surfaces. We propose a method to construct a classification Convolutional Neural Network (CNN) model based on the pre-trained U-Net, which is a well-known semantic segmentation model. Firstly, we train the U-Net with a limited amount of the asphalt pavement surface dataset which is obtained by a Mobile Mapping System (MMS). Then, we use the encoder of the trained U-Net as a feature extractor to construct a classification model, and train by fine-tuning. We describe comparative evaluations with VGG11, ResNet18, and GoogLeNet as well-known models constructed by transfer learning using ImageNet, which is a large size dataset of natural images. Experimental results show our model has high classification performance, compared to the other models constructed by transfer learning using ImageNet. Our method is effective to construct convolutional neural network model using the limited training dataset.

Humans predict the forest, not the trees: statistical learning of spatiotemporal structure in visual scenes.

The human brain is capable of using statistical regularities to predict future inputs. In the real world, such inputs typically comprise a collection of objects (e.g. a forest constitutes numerous trees). The present study aimed to investigate whether perceptual anticipation relies on lower-level or higher-level information. Specifically, we examined whether the human brain anticipates each object in a scene individually or anticipates the scene as a whole. To explore this issue, we first trained participants to associate co-occurring objects within fixed spatial arrangements. Meanwhile, participants implicitly learned temporal regularities between these displays. We then tested how spatial and temporal violations of the structure modulated behavior and neural activity in the visual system using fMRI. We found that participants only showed a behavioral advantage of temporal regularities when the displays conformed to their previously learned spatial structure, demonstrating that humans form configuration-specific temporal expectations instead of predicting individual objects. Similarly, we found suppression of neural responses for temporally expected compared with temporally unexpected objects in lateral occipital cortex only when the objects were embedded within expected configurations. Overall, our findings indicate that humans form expectations about object configurations, demonstrating the prioritization of higher-level over lower-level information in temporal expectation.

Subjective perception of objects depends on the interaction between the validity of context-based expectations and signal reliability

Predictive coding theories of visual perception assume that expectations based on prior knowledge modulate the processing of information. However, the underlying mechanisms remain debated. Some accounts propose that expectations enhance the perception of expected relative to unexpected stimuli while others assume the opposite. Recently, the opposing process theory suggested that enhanced perception of expected vs. unexpected stimuli may occur alternatively depending upon the reliability of the visual signal. When the signal is noisy, perception would be biassed toward what is expected since anything else may be too noisy to be resolved. When the signal is unambiguous, perception would be biassed toward what diverges from expectations and is more informative. Our study tested this hypothesis, using a perceptual matching task to investigate the influence of expectations on the perceived sharpness of objects in context. Participants saw two blurred images depicting the same object and had to adjust the blur level of one object to match the blur level of the other one. We manipulated the validity of expectations about objects by varying their scene context (congruent or incongruent context leading to valid or invalid expectations about the object). We also manipulated the reliability of the visual signal by varying the initial blur level of object pairs. Results showed that expectations validity differentially affected the perception of objects depending on signal reliability. Perception of validly expected objects was enhanced (sharpened) relative to unexpected objects when visual inputs were unreliable while this effect reversed to the benefit of unexpected objects when the signal was more reliable.

Inattentional blindness: Attentional set for efficient task success

Inattentional blindness is the failure to notice an unexpected object in plain sight when attention is otherwise engaged. We investigated what determines observers’ attentional set in a dynamic-counting inattentional blindness paradigm, when task instructions and visual distinctiveness of task-relevant objects were either congruent or in opposition. In seven experiments, observers counted bounces by task-relevant objects, with the instruction either to count-by-shape (squares, diamonds, crosses) or count-by-colour (blue, purple). To manipulate visual distinctiveness, we varied the extent to which task-relevant and task-irrelevant objects looked different on two dimensions: shape and colour. When colour better distinguished task-relevant from task-irrelevant objects, observers—even if instructed count-by-shape—reported an unexpected object that matched the colour of task-relevant objects. Crucially, when instructed count-by-colour, but shape better distinguished task-relevant from task-irrelevant objects, observers reported an unexpected object that matched the shape of task-relevant objects. We conclude that observers set their attention to promote efficient task performance.

Separation Axioms on Bipolar Hypersoft Topological Spaces

According to its definition, a topological space could be a highly unexpected object. There are spaces (indiscreet space) which have only two open sets: the empty set and the entire space. In a discrete space, on the other hand, each set is open. These two artificial extremes are very rarely seen in actual practice. Most spaces in geometry and analysis fall somewhere between these two types of spaces. Accordingly, the separation axioms allow us to say with confidence whether a topological space contains a sufficient number of open sets to meet our needs. To this end, we use bipolar hypersoft (BHS) sets (one of the efficient tools to deal with ambiguity and vagueness) to define a new kind of separation axioms called BHS Ti-space (i = 0, 1, 2, 3, 4). We show that ee BHS Ti-space (i = 1,2) implies BHS Ti−1-space; however, the converse is false, as shown by an example. e For i = 0, 1, 2, 3, 4, we prove that BHS Ti -space is hypersoft (HS) Ti -space and we present a condition so that eee HS Ti-space is BHS Ti-space. Moreover, we study that a BHS subspace of a BHS Ti-space is a BHS Ti-space for i = 0,1,2,3.

Performance Analysis of Localization Algorithms for Inspections in 2D and 3D Unstructured Environments Using 3D Laser Sensors and UAVs

One of the most relevant problems related to Unmanned Aerial Vehicle’s (UAV) autonomous navigation for industrial inspection is localization or pose estimation relative to significant elements of the environment. This paper analyzes two different approaches in this regard, focusing on its application to unstructured scenarios where objects of considerable size are present, such as a truck, a wind tower, an airplane, a building, etc. The presented methods require a previously developed Computer-Aided Design (CAD) model of the main object to be inspected. The first approach is based on an occupancy map built from a horizontal projection of this CAD model and the Adaptive Monte Carlo Localization (AMCL) algorithm to reach convergence by considering the likelihood field observation model between the 2D projection of 3D sensor data and the created map. The second approach uses a point cloud prior map of the 3D CAD model and a scan-matching algorithm based on the Iterative Closest Point Algorithm (ICP) and the Unscented Kalman Filter (UKF). The presented approaches have been extensively evaluated using simulated as well as previously recorded real flight data. We focus on aircraft inspection as a test example, but our results and conclusions can be directly extended to other applications. To support this assertion, a truck inspection has been performed. Our tests reflected that creating a 2D or 3D map from a standard CAD model and using a 3D laser scan on the created maps can optimize the processing time, resources and improve robustness. The techniques used to segment unexpected objects in 2D maps improved the performance of AMCL. In addition, we showed that moving around locations with relevant geometry after take-off when running AMCL enabled faster convergence and high accuracy. Hence, it could be used as an initial position estimation method for other localization algorithms. The ICP-NL method works well in environments with elements other than the object to inspect, but it can provide better results if some techniques to segment the new objects are applied. Furthermore, the proposed ICP-NL scan-matching method together with UKF performed faster, in a more robust manner, than NDT. Moreover, it is not affected by flight height. However, ICP-NL error may still be too high for applications requiring increased accuracy.

Typicality modulates the visual awareness of objects

In the present study, we assessed whether typicality can influence the visual awareness of objects. Participants tracked moving images of objects and counted how often members of one category bounced off the edges of the display. On the last trial, an unexpected object moved across the display. In our first two experiments, this object could belong to the same category as the tracked or untracked objects. While participants were more likely to notice atypical members of the untracked category, this pattern of results reversed when participants tracked atypical objects. In our last two experiments, the unexpected object could belong to the same category as the tracked objects or a new category of objects. In this case, participants were more likely to notice typical members of both the tracked category and the new category. Together, these findings suggest that typicality can modulate the visual awareness of objects.

Active inference through whiskers

Rodents use whisking to probe actively their environment and to locate objects in space, hence providing a paradigmatic biological example of active sensing. Numerous studies show that the control of whisking has anticipatory aspects. For example, rodents target their whisker protraction to the distance at which they expect objects, rather than just reacting fast to contacts with unexpected objects. Here we characterize the anticipatory control of whisking in rodents as an active inference process. In this perspective, the rodent is endowed with a prior belief that it will touch something at the end of the whisker protraction, and it continuously modulates its whisking amplitude to minimize (proprioceptive and somatosensory) prediction errors arising from an unexpected whisker–object contact, or from a lack of an expected contact. We will use the model to qualitatively reproduce key empirical findings about the ways rodents modulate their whisker amplitude during exploration and the scanning of (expected or unexpected) objects. Furthermore, we will discuss how the components of active inference model can in principle map to the neurobiological circuits of rodent whisking.