Search Results (223)

To understand human–machine interactions, we adopted a framework that distinguishes between stable individual differences (enduring personality/interpersonal traits), initial trust (pre-interaction expectations), and situational trust (dynamic calibration via gaze and behavior). A driving simulator experiment was conducted with 30 male participants to investigate trust calibration across three levels: manual (Level 0), semi-automated (Level 2, requiring monitoring), and fully automated (Level 4, system handles tasks). We combined eye tracking (pupillometry/fixations) with the Eysenck Personality Questionnaire (EPQ) and Interpersonal Trust Scale (ITS). Results indicated that semi-automation yielded a higher hazard detection sensitivity (d′ = 0.81) but induced greater physiological costs (pupil diameter, ηp² = 0.445) compared to manual driving. A mediation analysis confirmed that neuroticism was associated with initial trust specifically through interpersonal trust. Critically, despite lower initial trust, young male individuals with high interpersonal trust exhibited slower reaction times in the semi-automation model (B = 0.60, p = 0.035), revealing a “social complacency” effect where social faith paradoxically predicted lower behavioral readiness. Based on these findings, we propose that situational trust is a multi-layer calibration process involving dissociated attentional and behavioral mechanisms, suggesting that such “wary but complacent” drivers require adaptive HMI interventions. Full article

Optokinetic nystagmus (OKN) is an involuntary sawtooth eye movement that occurs in the presence of a drifting stimulus. Our experience is that low-amplitude/short-duration OKN can challenge the limits of our commercially available Pupil Neon eye-tracker, leading to false negative OKN detection results. We sought to investigate whether such instances could be remediated. We compared automated OKN detection using: (1) the gaze signal from the Pupil Neon (OKN-G), (2) centroid tracking (OKN-C), and (3) an image-phase-based “motion microscopy” technique (OKN-MMIC). The OKN-C and OKN-MMIC methods were also tested as a remediated step after a negative OKN-G result (OKN-C-STEP, OKN-MMIC-STEP). To validate the approaches adults (n = 22) with normal visual acuity was measured whilst viewing trials of an OKN induction stimulus shown at four levels of visibility. Confusion matrices and performance measures were determined for a “main” dataset that included all methods, and a “retest” set, which contained instances where centroid tracking failed. For the main set, all tested methods improved upon OKN-G by Matthew’s correlation coefficient (0.80–0.85 vs. 0.76), sensitivity (0.89–0.95 vs. 0.85), and accuracy (0.91–0.93 vs. 0.88); but only OKN-C yielded better specificity (0.90–0.96 vs. 0.95). For the retest set, MMIC and MMIC-STEP methods consistently improved upon the performance of OKN-G across all measures. Full article

This paper presents a novel hybrid multimodal deep learning model for robust and real-time eye gaze estimation. Accurate gaze tracking is essential for advancing human–computer interaction (HCI) and assistive technologies, but existing methods often struggle with environmental variations, require extensive calibration, and are computationally intensive. Our proposed model, GazeNet-HM, addresses these limitations by synergistically fusing features from RGB, depth, and infrared (IR) imaging modalities. This multimodal approach allows the model to leverage complementary information: RGB provides rich texture, depth offers invariance to lighting and aids pose estimation, and IR ensures robust pupil detection. Furthermore, we introduce a personalized adaptation module that dynamically fine-tunes the model to individual users with minimal calibration data. To ensure practical deployment, we employ advanced model compression techniques, enabling real-time inference on resource-constrained embedded systems. Extensive evaluations on public datasets (MPIIGaze, EYEDIAP, Gaze360) and our collected M-Gaze dataset demonstrate that GazeNet-HM achieves state-of-the-art performance, reducing the mean angular error by up to 27.1% compared to leading unimodal methods. After model compression, the system achieves a real-time inference speed of 32 FPS on an embedded Jetson Xavier NX platform. Ablation studies confirm the contribution of each modality and component, highlighting the effectiveness of our holistic design. Full article

Gaze behavior is a critical component of social interaction, reflecting emotional recognition and social regulation. While previous research has emphasized either situational influences (e.g., deception) or stable individual differences (e.g., attachment avoidance) on gaze patterns, studies exploring how these factors interact to shape gaze behavior in interpersonal contexts remain scarce. In this vein, the aim of the present study was to experimentally determine whether the gaze direction of individuals differs, with respect to their avoidant orientation, under changing situational conditions, including truthful and deceptive communication towards a counterpart. Using a within-person experimental design and the eye-tracking methodology, 31 participants took part in both rehearsed and spontaneous truth-telling and lie-telling tasks. Consistent with expectations, higher attachment avoidance was associated with significantly fewer fixations on emotionally expressive facial regions (e.g., mouth, jaw), and non-significant but visually consistent increases in fixations on the upper face (e.g., eyes) and background. These findings indicate that stable dispositional tendencies, rather than situational demands such as deception, predominantly shape gaze allocation during interpersonal interactions. They further provide a foundation for future investigations into the dynamic interplay between personality and situational context in interactive communicative settings. Full article

Background/Objectives: Mild cognitive impairment (MCI) affects multiple functional and cognitive domains, rendering it challenging to diagnose. Brief mental status exams are insensitive while detailed neuropsychological testing is time-consuming and presents accessibility issues. By contrast, the Oculo-Cognitive Addition Test (OCAT) is a rapid, objective tool that measures oculometric features during mental addition tasks under one minute. This study aims to develop artificial intelligence (AI)-derived predictive models using OCAT eye movement and time-based features for the early detection of those at risk for MCI, requiring more thorough assessment. Methods: The OCAT with integrated eye tracking was completed by 250 patients at the Mayo Clinic Arizona Department of Neurology. Raw gaze data analysis yielded time-related and eye movement features. Random Forest and univariate decision trees were the feature selection methods used to identify predictors of Dementia Rating Scale (DRS) outcomes. Logistic regression (LR) and K-nearest neighbors (KNN) supervised models were trained to classify PMCI using three feature sets: time-only, eye-only, and combined. Results: LR models achieved the highest performance using the combined time and eye movement features, with an accuracy of 0.97, recall of 0.91, and an AUPRC of 0.95. The eye-only and time-only LR models also performed well (accuracy = 0.93), though with slightly lower F1-scores (0.87 and 0.86, respectively). Overall, models leveraging both time and eye movement features consistently outperformed those using individual feature sets. Conclusions: Machine learning models trained on OCAT-derived features can reliably predict DRS outcomes (PASS/FAIL), offering a promising approach for early MCI identification. With further refinement, OCAT has the potential to serve as a practical and scalable cognitive screening tool, suitable for use in clinics, at the bedside, or in remote and resource-limited settings. Full article

Eye tracking (ET) technology is increasingly used in both research and clinical practice, but its accuracy may be compromised by the presence of ophthalmic lenses. This study systematically evaluated the influence of different optical prescriptions and lens treatments on ET performance using DIVE (Device for an Integral Visual Examination). Fourteen healthy participants underwent oculomotor control tests under thirteen optical conditions: six with varying dioptric powers and six with optical filters, compared against a no-lens control. Key parameters analysed included angle error, fixation stability (bivariate contour ellipse area, BCEA), saccadic accuracy, number of data gaps, and proportion of valid frames. High-powered spherical lenses (+6.00 D and −6.00 D) significantly increased gaze angle error, and the negative lens also increased data gaps, while cylindrical lenses had a moderate effect. Among filters, the Natural IR coating caused the greatest deterioration in ET performance, reducing valid samples and increasing the number of gaps with data loss, likely due to interference with the infrared-based detection system. The lens with basic anti-reflective treatment (SV Org 1.5 AR) also showed some deterioration in interaction with the ET. Other filters showed minimal or no significant impact. These findings demonstrate that both high-powered prescriptions and certain lens treatments can compromise ET data quality, highlighting the importance of accounting for optical conditions in experimental design and clinical applications. Full article

Strong communication is central to the translation of breast cancer screening availability into uptake. This experiment tests the role of design features of screening advertisements in directing visual attention in screening-eligible women (≥40 years). To this end, a within-subjects eye-tracking experiment (N = 30) was conducted in which women viewed six static public service advertisements. Predefined Areas of Interest (AOIs), Text, Image/Visual, Symbol, Logo, Website/CTA, and Source/Authority—were annotated, and three standard measures were calculated: Time to First Fixation (TTFF), Fixation Count (FC), and Fixation Duration (FD). Analyses combined descriptive summaries with subgroup analyses using nonparametric methods and generalized linear mixed models (GLMMs) employing participant-level random intercepts. Within each category of stimuli, detected differences were small in magnitude yet trended towards few revisits in each category for the FC mode; TTFF and FD showed no significant differences across categories. Viewing data from the perspective of Areas of Interest (AOIs) highlighted pronounced individual differences. Narratives/efficacy text and dense icon/text callouts prolonged processing times, although institutional logos and abstract/anatomical symbols generally received brief treatment except when coupled with action-oriented communication triggers. TTFF timing also tended toward individual areas of interest aligned with the Scan-Then-Read strategy, in which smaller labels/sources/CTAs are exploited first in comparison with larger headlines/statistical text. Practically, screening messages should co-locate access and credibility information in early-attention areas and employ brief, fluent efficacy text to hold gaze. The study adds PSA-specific eye-tracking evidence for breast cancer screening and provides immediately testable design recommendations for programs in Greece and the EU. Full article

Background: Early detection of autism spectrum disorder (ASD) is essential, as the first two years of life represent a critical window of neuroplasticity during which timely interventions can improve developmental outcomes. Traditional diagnostic methods, such as ADOS and ADI-R, rely on caregiver reports and structured observations, limiting ecological validity and accessibility. Eye-tracking (ET) offers a non-invasive, scalable approach to assess early atypical gaze patterns. Objectives: This systematic review and meta-analysis synthesized evidence on the diagnostic accuracy of ET for early ASD detection and its potential as an adjunctive screening tool. Methods: A comprehensive search of PubMed, Scopus, Web of Science, Medline, and the Cochrane Library identified studies published between January 2015 and July 2025. Eligible studies evaluated ET in infants and toddlers (≤36 months) for early ASD identification, following PRISMA guidelines. Results: Out of 513 records, 57 studies were included. Most studies reported reduced fixation on social stimuli, atypical gaze following, and preference for geometric over social images in infants later diagnosed with ASD. Pooled effect sizes indicated a moderate-to-large difference between ASD and typically developing groups in social fixation time (Hedges’ g ≈ 0.65, 95% CI: 0.48–0.82, I² = 58%). Studies integrating machine learning algorithms (n = 14) achieved improved sensitivity (up to 89%) and specificity (up to 86%) compared with conventional gaze metrics. Conclusions: Overall, ET shows strong potential as an early adjunctive screening method for ASD. Nonetheless, methodological heterogeneity and lack of standardized protocols currently limit clinical translation, underscoring the need for multi-center validation and task standardization. Full article

Early detection of Autism Spectrum Disorder (ASD), a neurodevelopmental condition characterized by social communication challenges, is essential for timely intervention. Naturalistic home videos collected via mobile applications offer scalable opportunities for digital diagnostics. We leveraged GuessWhat, a mobile game designed to engage parents and children, which has generated over 3000 structured videos from 382 children. From this collection, we curated a final analytic sample of 688 feature-rich videos centered on a single dyad, enabling more consistent modeling. We developed a two-step pipeline: (1) filtering to isolate high-quality videos, and (2) feature engineering to extract interpretable behavioral signals. Unimodal LSTM-based models trained on eye gaze, head position, and facial expression achieved test AUCs of 86% (95% CI: 0.79–0.92), 78% (95% CI: 0.69–0.86), and 67% (95% CI: 0.55–0.78), respectively. Late-stage fusion of unimodal outputs significantly improved predictive performance, yielding a test AUC of 90% (95% CI: 0.84–0.95). Our findings demonstrate the complementary value of distinct behavioral channels and support the feasibility of using mobile-captured videos for detecting clinically relevant signals. While further work is needed to improve generalizability and inclusivity, this study highlights the promise of real-time, scalable autism phenotyping for early interventions. Full article

Eye-tracking for user experience analysis has traditionally relied on dedicated hardware, which is often costly and imposes restrictive operating conditions. As an alternative, solutions utilizing ordinary webcams have attracted significant interest due to their affordability and ease of use. However, a major limitation persists in these vision-based methods: sensitivity to head movements. Therefore, users are often required to maintain a rigid head position, leading to discomfort and potentially skewed results. To address this challenge, this paper proposes a robust eye-tracking methodology designed to accommodate head motion. Our core technique involves mapping the displacement of the pupil center from a dynamically updated reference point to estimate the gaze point. When head movement is detected, the system recalculates the head-pointing coordinate using estimated head pose and user-to-screen distance. This new head position and the corresponding pupil center are then established as the fresh benchmark for subsequent gaze point estimation, creating a continuous and adaptive correction loop. We conducted accuracy tests with 22 participants. The results demonstrate that our method surpasses the performance of many current methods, achieving mean gaze errors of 1.13 and 1.37 degrees in two testing modes. Further validation in a smooth pursuit task confirmed its efficacy in dynamic scenarios. Finally, we applied the method in a real-world gaming context, successfully extracting fixation counts and gaze heatmaps to analyze visual behavior and UX across different game modes, thereby verifying its practical utility. Full article

Eye movement is an important tool used to investigate cognition. It also serves as input in human–computer interfaces for assistive technology. It can be measured with camera-based eye tracking and electro-oculography (EOG). EOG does not rely on eye visibility and can be measured even when the eyes are closed. We investigated the feasibility of detecting the gaze direction using EOG while having the eyes closed. A total of 15 participants performed a proprioceptive calibration task with open and closed eyes, while their eye movement was recorded with a camera-based eye tracker and with EOG. The calibration was guided by the participants’ hand motions following a pattern of felt dots on cardboard. Our cross-correlation analysis revealed reliable temporal synchronization between gaze-related signals and the instructed trajectory across all conditions. Statistical comparison tests and equivalence tests demonstrated that EOG tracking was statistically equivalent to the camera-based eye tracker gaze direction during the eyes-open condition. The camera-based eye-tracking glasses do not support tracking with closed eyes. Therefore, we evaluated the EOG-based gaze estimates during the eyes-closed trials by comparing them to the instructed trajectory. The results showed that EOG signals, guided by proprioceptive cues, followed the instructed path and achieved a significantly greater accuracy than shuffled control data, which represented a chance-level performance. This demonstrates the advantage of EOG when camera-based eye tracking is infeasible, and it paves the way for the development of eye-movement input interfaces for blind people, research on eye movement direction when the eyes are closed, and the early detection of diseases. Full article

Drowsy driving is a major cause of traffic accidents worldwide, and its early detection remains essential for road safety. Conventional driver monitoring systems (DMS) primarily rely on behavioral indicators such as eye closure, gaze, or head pose, which typically appear only after a significant decline in alertness. This study explores the potential of facial near-infrared (NIR) imaging as a hypothetical physiological indicator of drowsiness. Because NIR light penetrates more deeply into biological tissue than visible light, it may capture subtle variations in blood flow and oxygenation near superficial vessels. Based on this hypothesis, we conducted a pilot feasibility study involving young adult participants to investigate whether drowsiness levels could be estimated from single-frame NIR facial images acquired at 940 nm—a wavelength already used in commercial DMS and suitable for both physiological sensitivity and practical feasibility. A convolutional neural network (CNN) was trained to classify multiple levels of drowsiness, and Gradient-weighted Class Activation Mapping (Grad-CAM) was applied to interpret the discriminative regions. The results showed that classification based on 940 nm NIR images is feasible, achieving an optimal accuracy of approximately 90% under the binary classification scheme (Pattern A). Grad-CAM revealed that regions around the nasal dorsum contributed to this, consistent with known physiological signs of drowsiness. These findings support the feasibility of NIR-based drowsiness classification in young drivers and provide a foundation for future studies with larger and more diverse populations. Full article

We present the eyeglass-type switch, an eyeglass-mounted eye/blink switch designed for nurse-call operation by people with severe motor impairments, with a particular focus on amyotrophic lateral sclerosis (ALS). The system targets real-world bedside constraints—low illumination at night, supine posture, and network-independent operation—by combining near-infrared (NIR) LED illumination with an NIR eye camera and executing all processing on a small, GPU-free computer. A two-stage convolutional pipeline estimates eight periocular landmarks and the pupil center; eye-closure is detected either by a binary classifier or by an angle criterion derived from landmarks, which also skips pupil estimation during closure. User intent is determined by crossing a caregiver-tunable “off-area” around neutral gaze, implemented as rectangular or sector shapes. Four output modes—single, continuous, long-press, and hold-to-activate—are supported for both oculomotor and eyelid inputs. Safety is addressed via relay-based electrical isolation from the nurse-call circuit and audio feedback for state indication. The prototype runs at 18 fps on commodity hardware. In feature-point evaluation, mean errors were 2.84 pixels for landmarks and 1.33 pixels for the pupil center. In a bedside task with 12 healthy participants, the system achieved $F=0.965$ in single mode and $F=0.983$ in hold-to-activate mode; blink-only input yielded $F=0.993$. Performance was uniformly high for right/left/up and eye-closure cues, with lower recall for downward gaze due to eyelid occlusion, suggesting camera placement or threshold tuning as remedies. The results indicate that the proposed switch provides reliable, low-burden nurse-call control under nighttime conditions and offers a practical input option for emergency alerts and augmentative and alternative communication (AAC) workflows. Full article

This study presents a Digital Twin–Mixed Reality (DT–MR) framework for the immersive and interactive supervision of automated container terminals (ACTs), addressing the fragmented data and limited situational awareness of conventional 2D monitoring systems. The framework employs a middleware-centric architecture that integrates heterogeneous subsystems—covering terminal operation, equipment control, and information management—through standardized industrial communication protocols. It ensures synchronized timestamps and delivers semantically aligned, low-latency data streams to a multi-scale Digital Twin developed in Unity. The twin applies level-of-detail modeling, spatial anchoring, and coordinate alignment (from Industry Foundation Classes (IFCs) to east–north–up (ENU) coordinates and Unity space) for accurate registration with physical assets, while a Microsoft HoloLens 2 device provides an intuitive Mixed Reality interface that combines gaze, gesture, and voice commands with built-in safety interlocks for secure human–machine interaction. Quantitative performance benchmarks—latency ≤100 ms, status refresh ≤1 s, and throughput ≥10,000 events/s—were met through targeted engineering and validated using representative scenarios of quay crane alignment and automated guided vehicle (AGV) rerouting, demonstrating improved anomaly detection, reduced decision latency, and enhanced operational resilience. The proposed DT–MR pipeline establishes a reproducible and extensible foundation for real-time, human-in-the-loop supervision across ports, airports, and other large-scale smart infrastructures. Full article

Eye movements are regulated by the ocular motor nerves (cranial nerves [CNs] III, IV, and VI), which control the six extraocular muscles of each eye. Palsies of CNs III, IV, and VI can restrict eye movements, resulting in strabismus and diplopia, and so clinical evaluations of eye movements are crucial for diagnosing CN palsies. This study aimed to develop an accurate artificial intelligence (AI) system for classifying CN III, IV, and VI palsies using nine-gaze ocular photographs. We analyzed 478 nine-gaze photographs comprising 70, 29, and 58 cases of CN III, IV, and VI palsies, respectively. The images were processed using MATLAB. For model training, each photograph of eye movements in the nine directions was numerically coded. A multinetwork model was employed to ensure precise analyses of paralytic strabismus. The AI system operates by referring data on minor abnormalities in the nine-gaze image to a network designed to detect CN IV abnormalities, which re-examines downward and lateral gazes to detect distinctions. Data on major abnormalities are directed to a different network trained to differentiate between CN III and VI abnormalities. EfficientNet-B0 was applied to reduce overfitting and improve learning efficiency in training with limited medical imaging data as the neural network architecture. The diagnostic accuracies of the proposed network for CN III, IV, and VI palsies were 99.31%, 97.7%, and 98.22%, respectively. This study has demonstrated the design of an AI model using a relatively small dataset and a multinetwork training system for analyzing nine-gaze photographs in strabismus patients with CN III, IV, and VI palsies, achieving an overall accuracy of 98.77%. Full article