Non-face Images Research Articles

Dynamic brain communication underwriting face pareidolia

Face pareidolia is a tendency to seeing faces in nonface images that reflects high tuning to a face scheme. Yet, studies of the brain networks underwriting face pareidolia are scarce. Here, we examined the time course and dynamic topography of gamma oscillatory neuromagnetic activity while administering a task with nonface images resembling a face. Images were presented either with canonical orientation or with display inversion that heavily impedes face pareidolia. At early processing stages, the peaks in gamma activity (40 to 45 Hz) to images either triggering or not face pareidolia originate mainly from the right medioventral and lateral occipital cortices, rostral and caudal cuneus gyri, and medial superior occipital gyrus. Yet, the difference occurred at later processing stages in the high-frequency range of 80 to 85 Hz over a set of the areas constituting the social brain. The findings speak rather for a relatively late neural network playing a key role in face pareidolia. Strikingly, a cutting-edge analysis of brain connectivity unfolding over time reveals mutual feedforward and feedback intra- and interhemispheric communication not only within the social brain but also within the extended large-scale network of down- and upstream regions. In particular, the superior temporal sulcus and insula strongly engage in communication with other brain regions either as signal transmitters or recipients throughout the whole processing of face-pareidolia images.

Suppressing Uncertainty in Gaze Estimation

Uncertainty in gaze estimation manifests in two aspects: 1) low-quality images caused by occlusion, blurriness, inconsistent eye movements, or even non-face images; 2) uncorrected labels resulting from the misalignment between the labeled and actual gaze points during the annotation process. Allowing these uncertainties to participate in training hinders the improvement of gaze estimation. To tackle these challenges, in this paper, we propose an effective solution, named Suppressing Uncertainty in Gaze Estimation (SUGE), which introduces a novel triplet-label consistency measurement to estimate and reduce the uncertainties. Specifically, for each training sample, we propose to estimate a novel ``neighboring label'' calculated by a linearly weighted projection from the neighbors to capture the similarity relationship between image features and their corresponding labels, which can be incorporated with the predicted pseudo label and ground-truth label for uncertainty estimation. By modeling such triplet-label consistency, we can largely reduce the negative effects of unqualified images and wrong labels through our designed sample weighting and label correction strategies. Experimental results on the gaze estimation benchmarks indicate that our proposed SUGE achieves state-of-the-art performance.

Do subtle cultural differences sculpt face pareidolia?

Face tuning to non-face images such as shadows or grilled toasts is termed face pareidolia. Face-pareidolia images represent a valuable tool for investigation of social cognition in mental disorders. Here we examined (i) whether, and, if so, how face pareidolia is affected by subtle cultural differences; and (ii) whether this impact is modulated by gender. With this purpose in mind, females and males from Northern Italy were administered a set of Face-n-Thing images, photographs of objects such as houses or waves to a varying degree resembling a face. Participants were presented with pareidolia images with canonical upright orientation and display inversion that heavily affects face pareidolia. In a two-alternative forced-choice paradigm, beholders had to indicate whether each image resembled a face. The outcome was compared with the findings obtained in the Southwest of Germany. With upright orientation, neither cultural background nor gender affected face pareidolia. As expected, display inversion generally mired face pareidolia. Yet, while display inversion led to a drastic reduction of face impression in German males as compared to females, in Italians, no gender differences were found. In a nutshell, subtle cultural differences do not sculpt face pareidolia, but instead affect face impression in a gender-specific way under unusual viewing conditions. Clarification of the origins of these effects requires tailored brain imaging work. Implications for transcultural psychiatry, in particular, for schizophrenia research, are highlighted and discussed.

Assessing the perception of face pareidolia in children (Homo sapiens), rhesus monkeys (Macaca mulatta), and capuchin monkeys (Sapajus apella).

Face pareidolia is the misperception of a face in an inanimate object and is a common feature of the face detection system in humans. Whereas there are many similarities in how humans and nonhuman animals such as monkeys perceive and respond to faces, it is still unclear whether other species also perceive certain nonface stimuli as faces. We presented a novel computerized task to capuchin monkeys (Sapajus apella), rhesus monkeys (Macaca mulatta), and preschool-aged children (Homo sapiens). This task trained subjects to choose faces over nonface images, and then presented pareidolia images with nonface images. All species selected faces most often on trials that included face images. However, only children selected pareidolia images at levels above chance. These results indicate that while children report perceiving face pareidolia, monkeys do not. These species differences could be due to human-unique experiences that result in an increased aptitude for anthropomorphizing objects with face-like patterns. It could also be due to monkeys showing a greater reliance on stimulus features rather than global, holistically organized cues that faces provide. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Face pareidolia in male schizophrenia

Faces are valuable signals for efficient social interaction. Yet, social cognition including the sensitivity to a coarse face scheme may be deviant in schizophrenia (SZ). Tuning to faces in non-face images such as shadows, grilled toasts, or ink blots is termed face pareidolia. This phenomenon is poorly investigated in SZ. Here face tuning was assessed in 44 male participants with SZ and person-by-person matched controls by using recently created Face-n-Thing images (photographs of non-face objects to a varying degree resembling a face). The advantage of these images is that single components do not automatically trigger face processing. Participants were administered a set of images with upright and inverted (180° in the image plane) orientation. In a two-alternative forced-choice paradigm, they had to indicate whether an image resembled a face. The findings showed that: (i) With upright orientation, SZ patients exhibited deficits in face tuning: they provided much fewer face responses than controls. (ii) Inversion generally hindered face pareidolia. However, while in neurotypical males, inversion led to a drastic drop in face impression, in SZ, the impact of orientation was reduced. (iii) Finally, in accord with the signal detection theory analysis, the sensitivity index (d-prime) was lower in SZ, whereas no difference occurred in decision criterion. The outcome suggests altered face pareidolia in SZ is caused by lower face sensitivity rather than by alterations in cognitive bias. Comparison of these findings with earlier evidence confirms that tuning to social signals is lower in SZ, and warrants tailored brain imaging research.

Deep convolutional network based real time fatigue detection and drowsiness alertness system

<span>Fatigue and drowsiness detection techniques based on the external features are under progress, and the methods of facial feature extraction require further development. This paper discusses the innovative processes, efficient methods, and recent advancements in the field of drowsiness and fatigue detection. In this proposed model, a wide application is planned in the field of artificial intelligence by defining the fundamentals of human-computer interaction, facial expression recognition and driver fatigue-sleepiness determination. This research outlines an efficient and effective three-phase strategy for detecting drowsiness. Viola Jones is used to detect facial traits in these three phases. Detection of yawning and tracking once the face has been identified, the segmenting the skin, the system becomes lighting invariant portion by itself, focusing on the chromatic components based on skin, and to reject most of non-face image backdrops. The color eye tracking and yawning detection are carried out by template matching with the correlation coefficient. The vectors of features based on each of the above phases is concatenated, and a binary result is obtained. The analysis of sound and successive frames into fatigue and non-fatigue states has been classified. If the time in fatigue state exceeds the threshold, the system will sound an alarm. </span>

Detection of duplicate and non-face images in the eRecruitment applications using machine learning techniques

The objective of this work is to develop methodologies to detect, and report the noncompliant images with respect to indian space research organisation (ISRO) recruitment requirements. The recruitment software hosted at U. R. rao satellite centre (URSC) is responsible for handling recruitment activities of ISRO. Large number of online applications are received for each post advertised. In many cases, it is observed that the candidates are uploading either wrong or non-compliant images of the required documents. By non-compliant images, we mean images which do not have faces or there is not enough clarity in the faces present in the images uploaded. In this work, we attempt to address two specific problems namely: 1) To recognise image uploaded to recruitment portal contains a human face or not. This is addressed using a face detection algorithm. 2) To check whether images uploaded by two or more applications are same or not. This is achieved by using machine learning (ML) algorithms to generate similarity score between two images, and then identify the duplicate images. Screening of valid applications becomes very challenging as the verification of such images using a manual process is very time consuming and requires large human efforts. Hence, we propose novel ML techniques to determine duplicate and non-face images in the applications received by the recruitment portal.

Explaining face representation in the primate brain using different computational models

Understanding how the brain represents the identity of complex objects is a central challenge of visual neuroscience. The principles governing object processing have been extensively studied in the macaque face patch system, a sub-network of inferotemporal (IT) cortex specialized for face processing. A previous study reported that single face patch neurons encode axes of a generative model called the "active appearance" model, which transforms 50D feature vectors separately representing facial shape and facial texture into facial images. However, a systematic investigation comparing this model to other computational models, especially convolutional neural network models that have shown success in explaining neural responses in the ventral visual stream, has been lacking. Here, we recorded responses of cells in the most anterior face patch anterior medial (AM) to a large set of real face images and compared a large number of models for explaining neural responses. We found that the active appearance model better explained responses than any other model except CORnet-Z, a feedforward deep neural network trained on general object classification to classify non-face images, whose performance it tied on some face image sets and exceeded on others. Surprisingly, deep neural networks trained specifically on facial identification did not explain neural responses well. A major reason is that units in the network, unlike neurons, are less modulated by face-related factors unrelated to facial identification, such as illumination.

Face sensitivity: Effects of gender and orientation

IntroductionResearch on face tuning is of particular relevance during the Covid-19 pandemic leading to social isolation and anxiety, but also requiring social integrity. Face sensitivity represents an essential component of social competence. This ability is aberrant in most neuropsychiatric conditions. Studies in typically developing individuals enable to develop new tools for examination and better understanding non-verbal social cognition in neuropsychiatry.ObjectivesHere we used a novel set of Face-n-Thing images to address the following issues: (i) whether the ability to seeing faces in non-face images (face pareidolia) is affected by gender; and (ii) whether it is altered with changing display orientation. The main advantage of Face-n-Thing images is that face tuning occurs without being explicitly fostered by familiar elements.MethodsA newly developed Face-n-Thing task, on which images were shown either with canonical upright orientation or inverted 180º in the image plane, was administered to healthy females and males. On each trial, they have to indicate whether they have a face impression.ResultsFace impression was substantially impeded by display inversion in both males and females. With upright display orientation, no gender differences were found, whereas with inversion, Face-n-Thing images elicited face impression in females significantly more often.ConclusionsThe findings open a way for examination of face sensitivity and underwriting brain networks in neuropsychiatric conditions, most of which are gender-specific. Display inversion represents a proper control for face tuning in neuroimaging studies. Gender differences should be taken into account when conceiving studies in neuropsychiatric populations.DisclosureNo significant relationships.

Face tuning in female and male individuals with depression

IntroductionThe current COVID-19 pandemic brings social isolation to our daily lives that may elevate depression. The impact of major depressive disorder (MDD) on social cognitive functioning is far from understood, but essential for prevention and treatment of this neuropsychiatric condition.ObjectivesOur aim was to examine (i) whether face tuning is lower in depression; and (ii) how it is related to other cognitive abilities (such as perceptional organization). Furthermore, we intended to clarify gender impact on face tuning in MDD, as twice more females are affected.MethodsUsing a recently developed paradigm, the Face-n-Food task, we examined face tuning in 26 patients with MDD and 26 person-by-person matched controls. The advantage of non-face images is that its single elements do not promote face processing.ResultsStrikingly, MDD individuals showed intact face tuning. As sex ratio in our patient sample was about 2:1 (as in MDD population in general), we recruited additional male patients and found that MDD male patients were as good as female patients. Yet, while face tuning in MDD patients showed a significant correlation with perceptual organization abilities, in controls, it was linked with social cognition.ConclusionsThe outcome suggests that the origins of aberrant social functioning in MDD lie in maladaptive cognitive schemas rather than in a lack of sensitivity towards social signals per se. To elucidate neural circuits involved in face tuning in MDD, a magnetoencephalography (MEG) study with the Face-n-Food images is currently under progress.DisclosureNo significant relationships.

Joint estimation of low-rank components and connectivity graph in high-dimensional graph signals: Application to brain imaging

Given high-dimensional, graph-smooth and grossly-corrupted signals, this paper presents the algorithm to simultaneously estimate the intrinsic low-rank components and the underlying graph from high-dimensional noisy graph signal. We assume that the perturbation on low-rank components is sparse and signals are smooth on unknown underlying graph. The proposed algorithm learns the low-rank components by exploiting estimate of the graph, and refines the graph estimation with learned low-rank components. We propose two solutions to this problem: One applies alternating optimization to solve the subproblems. The other solves the problem directly. Furthermore, we analyze the impact of inexact graph on low-rank components estimation to justify our approach. We conduct extensive experiments on the proposed algorithm with synthetic data and real brain imaging data, Magnetoencephalography (MEG) and compare it with state-of-the-art methods. In particular, our algorithm is applied to estimate the low-rank components for classifying MEG signals evoked when a subject views a face or non-face image. We observe that our proposed algorithm is competitive in estimating low-rank components, adequately capturing intrinsic task-related information in lower-dimensional representation. This leads to improved classification performance. Furthermore, we notice that our estimated graph indicates brain active regions for the visual activity that are consistent with neuroscientific findings.

Neural responses in a fast periodic visual stimulation paradigm reveal domain-general visual discrimination deficits in developmental prosopagnosia

We investigated selective impairments of visual identity discrimination in developmental prosopagnosia (DP), using a fast periodic identity oddball stimulation paradigm with electroencephalography (EEG). In Experiment 1, neural responses to unfamiliar face identity changes were strongly attenuated for individuals with DP as compared to Control participants, to the same extent for upright and inverted faces. This reduction of face identity discrimination responses, which was confirmed in Experiment 2, provides direct evidence for deficits in the visual processing of unfamiliar facial identity in DP. Importantly, Experiment 2 demonstrated that DPs showed attenuated neural responses to identity oddballs not only with face images, but also with non-face images (cars). This result strongly suggests that rapid identity-related visual processing impairments in DP are not restricted to faces, but also affect familiar classes of non-face stimuli. Visual discrimination deficits in DP do not appear to be face-specific. To account for these findings, we propose a new account of DP as a domain-general deficit in rapid visual discrimination.

Typical visual unfamiliar face individuation in left and right mesial temporal epilepsy

Patients with chronic mesial temporal lobe epilepsy have difficulties at identifying familiar faces as well as at explicit old/new face recognition tasks. However, the extent to which these difficulties can be attributed to visual individuation of faces, independently of general explicit learning and semantic memory processes, is unknown. We tested 42 mesial temporal lobe epilepsy patients divided into two groups according to the side of epilepsy (left and right) and 42 matched controls on an extensive series of individuation tasks of unfamiliar faces and control visual stimuli, as well as on face detection, famous face recognition and naming, and face and non-face learning. Overall, both patient groups had difficulties at identifying and naming famous faces, and at explicitly learning face and non-face images. However, there was no group difference in accuracy between patients and controls at the two most widely used neuropsychological tests assessing visual individuation of unfamiliar faces (Benton Facial Recognition Test and Cambridge Face Memory Test). While patients with right mesial temporal lobe epilepsy were slowed down at all tasks, this effect was not specific to faces or even high-level stimuli. Importantly, both groups showed the same profile of response as typical participants across various stimulus manipulations, showing no evidence of qualitative processing impairments. Overall, these results point to largely preserved visual face individuation processes in patients with mesial temporal lobe epilepsy, with semantic and episodic memory difficulties being consistent with the localization of the neural structures involved in their epilepsy (anterior temporal cortex and hippocampus). These observations have implications for the prediction of neuropsychological outcomes in the case of surgery and support the validity of intracranial electroencephalographic recordings performed in this population to understand neural mechanisms of human face individuation, notably through intracranial electrophysiological recordings and stimulations.

Face pareidolia in schizophrenia

BackgroundFaces convey valuable daily life social signals. As in most psychiatric conditions, non-verbal social cognition or its components including face processing may be aberrant in schizophrenia (SZ). Social participation of individuals with SZ is vital for their quality of life, and remediation of social abilities in this population is of high relevance both for society and clinical care. MethodTuning to faces in non-face images such as shadows, grilled toasts, or ink blots is called face pareidolia. Humans possess high sensitivity to facial signals: even fetuses and infants are well tuned to coarse face cues. Here we assessed face tuning in individuals with SZ and person-by-person matched controls by using a new experimental tool, a set of food-plate images bordering on the Giuseppe Arcimboldo style. The key benefit of these images is that single components do not trigger face processing. Results and conclusionsThe outcome indicates that individuals with SZ exhibit aberrant face tuning in face-like non-face images (χ2(1) = 17.44, p = 0.0001) that can hamper adaptive interaction with peers and social participation hindering, in turn, clinical remediation. Face response rate in SZ patients was related to the scores on the event arrangement task tapping social cognition (Pearson product-moment correlation, r = 0.602, p = 0.01) and on picture completion task assessing visual perceptual organization (Spearman's rho = 0.614, p = 0.009). Therefore, poor performance on the face tuning task is unlikely to be accounted for by deviant general cognitive abilities, but rather by impairments in perceptual integration and social cognition. Comparison of these findings with data in autism and other neuropsychiatric conditions provides novel insights on the origins of face tuning in SZ and triggers brain imaging research.

Baby Monitoring System using Image Processing and IoT

Non-contact-based baby monitoring system using image processing is proposed in this paper which is used for proper safety and monitoring the activity of baby by their busy parents. The system detects the motion, crying and present position of the baby. If any abnormal action is detected, then the system sends a message in the form of text and images of baby to the particular user through email. Raspberry Pi B+ module is used to process the videos taken by pi camera, MIC is used for crying detection and image processing is used for detection of real-time motion of babies and boundary condition of the bed. The system required to first install OS Raspbian, and all the other packages like OpenCV, Numpy and Virtual environment. Face detection algorithm is trained using Haar classifier for positive face images and negative nonface images. This system will help in decreasing the chances of the baby’s falling from the bed. Also, this system can be used in hospitals while baby is sleeping where the stress among the nurses will be reduced.

Maternal odor shapes rapid face categorization in the infant brain.

To successfully interact with a rich and ambiguous visual environment, the human brain learns to differentiate visual stimuli and to produce the same response to subsets of these stimuli despite their physical difference. Although this visual categorization function is traditionally investigated from a unisensory perspective, its early development is inherently constrained by multisensory inputs. In particular, an early-maturing sensory system such as olfaction is ideally suited to support the immature visual system in infancy by providing stability and familiarity to a rapidly changing visual environment. Here, we test the hypothesis that rapid visual categorization of salient visual signals for the young infant brain, human faces, is shaped by another highly relevant human-related input from the olfactory system, the mother's body odor. We observe that a right-hemispheric neural signature of single-glance face categorization from natural images is significantly enhanced in the maternal versus a control odor context in individual 4-month-old infant brains. A lack of difference between odor conditions for the common brain response elicited by both face and non-face images rules out a mere enhancement of arousal or visual attention in the maternal odor context. These observations show that face-selective neural activity in infancy is mediated by the presence of a (maternal) body odor, providing strong support for multisensory inputs driving category acquisition in the developing human brain and having important implications for our understanding of human perceptual development.

Predicting Blood Oxygenation Level-Dependent Activity in Fusiform Face Area from the Activity in Other Visual Areas.

Neuroimaging studies have shown that discrete regions in ventral visual pathway respond selectively to specific object categories. For example, the fusiform face area (FFA) in humans is consistently more responsive to face than nonface images. However, it is not clear how other cortical regions contribute to this preferential response in FFA. To address this question, we performed a functional magnetic resonance imaging study on human subjects watching naturalistic movie clips from human actions. We then used correlation and multivariate regression (partial least-squares regression) analyses to estimate/predict the mean BOLD (blood oxygenation level-dependent) activity in FFA, from the mean and pattern of responses in 24 visual cortical areas. Higher tier retinotopic areas V3, hV4, and LO2, motion-selective area middle temporal, body-selective areas, and non-FFA face-selective areas had the best prediction accuracy particularly when they were located ipsilateral to FFA. All non-FFA collectively could explain up to 75% of variance in the FFA response. The regression models were also designed to predict the mean activity in one face area from the pattern of activity in another face area. The prediction power was significantly higher between the occipital face area and FFA. The multivariate regression analysis provides a new framework for investigating functional connectivity between cortical areas, and it could inform hierarchical models of visual cortex.

Robust Jointly Sparse Regression with Generalized Orthogonal Learning for Image Feature Selection

Ridge regression (RR) and its variants are fundamental methods for multivariable data analysis, which have been widely used to deal with different problems in pattern recognition or classification. However, these methods have their common drawback. That is, the number of the learned projections is limited by the number of class. Moreover, most of these methods do not consider the local structure of the data, which makes them less competitive in the case when data are lying on a lower dimensional manifold. Therefore, in this paper, we propose a robust jointly sparse regression method to integrate the locality geometric structure with generalized orthogonality constraint and joint sparsity into a regression modal to address these problems. The optimization model can be solved by an alternatively iterative algorithm using orthogonal matching pursuit (OMP) and singular value decomposition. Experimental results on face and non-face image database demonstrate the superiority of the proposed method. The matlab code can be found at http://www.scholat.com/laizhihui.

A Novel Face Detection Method Based on Over-complete Incoherent Dictionary Learning

In this paper, face detection problem is considered using the concepts of compressive sensing technique. This technique includes dictionary learning procedure and sparse coding method to represent the structural content of input images. In the proposed method, dictionaries are learned in such a way that the trained models have the least degree of coherence to each other. The novelty of the proposed method involves the learning of comprehensive models with atoms that have the highest atom/data coherence with the training data and the lowest within-class and between-class coherence parameters. Each of these goals can be achieved through the proposed procedures. In order to achieve the desired results, a variety of features are extracted from the images and used to learn the characteristics of face and non-face images. Also, the results of the proposed classifier based on the incoherent dictionary learning technique are compared with the results obtained from the other common classifiers such as neural network and support vector machine. Simulation results, along with a significance statistical test show that the proposed method based on the incoherent models learned by the combinational features is able to detect the face regions with high accuracy rate.

Social Cognition in Down Syndrome: Face Tuning in Face-Like Non-Face Images.

Individuals with Down syndrome (DS) are widely believed to possess considerable socialization strengths. However, the findings on social cognition capabilities are controversial. In the present study, we investigated whether individuals with DS exhibit shortage in face tuning, one of the indispensable components of social cognition. For this purpose, we implemented a recently developed Face-n-Food paradigm with food-plate images composed of food ingredients such as fruits and vegetables. The key benefit of such face-like non-face images is that single elements do not facilitate face processing. In a spontaneous recognition task, 25 children with DS aged 9 to 18 years were presented with a set of Face-n-Food images bordering on the Giuseppe Arcimboldo style. The set of images was administered in a predetermined order from the least to most resembling a face. In DS individuals, thresholds for recognition of the Face-n-Food images as a face were drastically higher as compared not only with typically developing controls, but also with individuals with autistic spectrum disorders and Williams-Beuren syndrome. This outcome represents a significant step toward better conceptualization of the visual social world in DS and neurodevelopmental disorders in general.