搜索结果：Vision research

There has been a consistent increase in the global population of remote workers. Yet the linkage between occupational mental health and the most prevalent ergonomics-related stressor among remote workers, computer vision syndrome, has not received comprehensive research attention. Data were gathered from 254 participants. Data analysis was conducted using simple and hierarchical linear regression. Results revealed that computer vision syndrome has a significantly negative influence on occupational mental health (F(1,252) = 18.88, p < 0.001), even after controlling for age and gender (&#x3b2; = -0.27, t = -4.32, p < 0.001). It was also shown that high employee resilience has a direct, significant positive impact on occupational mental health (F(1,252) = 12.79, p < 0.001). Through interaction, a significant moderating effect of employee resilience in the relationship between computer vision syndrome and occupational mental health was observed (Model 1 [computer vision syndrome]: &#x3b2; = -0.26, t = -4.34, p < 0.001; Model 2 [computer vision syndrome&#x2009;&#xd7;&#x2009;employee resilience]: &#x3b2; = 0.34, t = 3.45, p < 0.001).

Artificial intelligence (AI) technologies for vision-based epilepsy monitoring are advancing rapidly in health care. Despite growing research using various video data sources and analytical approaches, no comprehensive framework exists to classify these technologies. This scoping review aimed to develop and validate a taxonomy for AI technologies in vision-based epilepsy monitoring and to characterize visual AI approaches in epilepsy care. Using an extended taxonomy development framework, we developed the taxonomy in 5 iterative cycles, drawing on theory and practice. We conducted a scoping review, market analysis, and applicability evaluation with market-ready solutions. We searched Scopus, Web of Science, and PubMed, including MeSH (Medical Subject Headings) terms; the final search was completed on January 16, 2026. We included primary studies from 2013 onward on AI-based or machine learning-based monitoring or prediction of epileptic seizures in humans using visual data. We excluded reviews, non-English publications, nonepilepsy studies, studies focused only on electroencephalography or wearables, animal studies, and pre-2013 publications. Evidence was charted through narrative and tabular synthesis and descriptive frequency analysis. In line with scoping review guidance, we did not conduct a meta-analysis or critical appraisal. To assess validity and practical relevance, 9 domain experts evaluated the taxonomy using a Delphi technique. We included 40 original studies. Study analysis yielded 16 dimensions, including data acquisition source, tracking target, image processing, classifier type, performance metrics, environment, seizure classification, data privacy, and user interface. Expert feedback added 4 further dimensions, including communication mode and information purpose. The final taxonomy comprises 23 dimensions with 102 characteristics. The review identified structural evidence gaps across settings, evaluation maturity, and reporting practices. Detection and classification in stationary settings predominated, whereas predictive approaches and real-time feedback were limited. Deep learning detection methods were common, but performance reporting was inconsistent, and patient-facing functionalities were limited. Privacy safeguards and standardized metrics were often incompletely reported, reducing comparability and maturity assessment. The taxonomy translates these patterns into guidance for benchmarking, procurement evaluation, user interface, and explainable AI design. We synthesized 5 main findings and 10 implications for research and practice. Key challenges concern standardization, seizure prediction, and real-time applicability. Vision-based AI technologies for epilepsy monitoring are still dominated by proof-of-concept and pilot evaluations, indicating a gap between technical feasibility and deployment-ready systems. This scoping review presents an implementation-oriented taxonomy integrating application context, system architecture, visual analysis, AI models, performance reporting, and feedback design into a single classification framework. Unlike prior work that mainly maps methods or data sources, the taxonomy provides a shared structure for consistent system-level characterization and comparison across studies and emerging solutions. It may support benchmarking, implementation-focused evaluation, procurement, and translation into clinical and home settings.

Reasoning is a hallmark of human intelligence, enabling adaptive decision-making in complex unfamiliar scenarios. In contrast, machine intelligence remains bound to training data, unable to dynamically refine solutions at inference. While recent advances have explored machine reasoning - trading inference-time compute for improved performance - they focus on verbal domains such as mathematical problem-solving where explicit rules govern step-by-step solution generation. Many tasks lack sufficient labelled data and require alternative performance improvement mechanisms, such as inference-time compute. Here we present a paradigm for machine reasoning in vision, enabling performance improvements with increasing thinking time (inference-time compute), even with limited labelled data. Our approach is inspired by dual-process theories of human cognition, integrating a fast-thinking System I module for generating and verifying solutions in familiar tasks, with a slow-thinking System II module that iteratively refines predictions using self-play reinforcement learning, even when task-specific data is limited. This paradigm involves proposing, competing over, and refining solutions until convergence. We demonstrate that extended inference-time compute yields superior performance compared to large-scale supervised learning, foundation models, and human experts in vision tasks. These include computer-vision benchmarks and cancer localisation across five organs, highlighting the potential of inference-time compute for data-scarce problems.

The dynamic vision sensor captures visual information as discrete events, enabling high-speed imaging with reduced data redundancy, but is limited by lack of color sensitivity, a speed-noise trade-off, and inefficient data transfer. Here we show an amphibian-inspired dynamic vision system (ADVS) based on ferroelectric field-effect transistors that emulates the hierarchical functions of amphibian retinas, including spectral perception, spatial preprocessing, and event-driven neural encoding. The ferroelectric transistors exhibit broadband photosensitivity (365-637&#x2009;nm) and bidirectional photoresponses, enabling multichannel spectral recognition from ultraviolet to visible light. Device arrays further reproduce center-surround receptive-field-like processing, enhancing spatial contrast while suppressing background noise under weak illumination. Owing to the steep switching characteristics (SSmin = 53.8&#x2009;mV dec-1) of the transistors, the system also supports microsecond-scale event-driven spiking responses. Combined with bioinspired hierarchical preprocessing framework and event-driven convolutional neural network, the ADVS achieves 96.5% accuracy in dynamic facial expression recognition and real-time multi-agent trajectory prediction with <5% error.

The purpose of this study was to investigate and quantify the influence of stimulus chromaticity on the perception of visual disturbances, specifically halos, under dim lighting conditions. The study also investigated age-related variations in this perception. Fifty healthy participants were divided into two age groups of 25 each: young adults (<25 years) and older adults (>54 years). The halo perception was quantified using the light disturbance analyser (LDA), a validated device designed to assess visual disturbances such as halos and glare under controlled lighting conditions. To assess the effect of stimulus chromaticity on halo perception, three filters with spectral transmittances centred in the red, green and blue regions of the visible spectrum were used. Measurements were recorded both before and after compensating for longitudinal chromatic aberration (LCA). In both age groups, white and green colours produced the smallest angular size of the perceived halo, followed by red, whereas blue induced the largest halo size. While LCA compensation under blue light was sufficient for the younger group to perceive a halo size similar to that under white light, this compensation proved insufficient for the older group. Perceived halo size was greatest when caused by a blue stimulus, followed by red light, while white and green sources yielded halos with comparable, smaller sizes across both age groups. The influence of age on perceived halo size under blue light was statistically significant. Furthermore, LCA compensation resulted in a greater benefit in perceived halo size for the younger, compared with the older group, under blue light.

Research on visually guided object manipulation has shown that participants fixate goal locations-including objects to be grasped and locations where they are placed-prior to hand arrival, with gaze serving two primary functions: directing the hand (or object in hand) to the vicinity of the goal using peripheral vision and gaze related signals, and guiding the hand using central vision as it approaches the goal. However, real world manipulation tasks are often performed while concurrent monitoring of the environment, resulting in competition for gaze. We examined gaze-hand coordination under such conditions. Human participants of either sex performed a manipulation task, that involved grasping balls and placing them at target locations, while concurrently monitoring a display to detect probabilistically occurring visual events, which required central vision. Participants managed gaze competition in two main ways. First, fixations allocated to the action task were brief and prioritized directing the hand towards the goal (object or target location); participants then relied on tactile feedback to complete the action (grasping or placing the object). When tactile feedback was reduced-by using a tool instead of the fingertips to perform the task-gaze additionally served the guiding function. Second, participants reduced gaze competition by exploiting temporal regularities of events in the monitoring task. Specifically, they adjusted both gaze allocation and hand movement timing to reduce the likelihood that action task fixations would coincide with visual events. These findings demonstrate how individuals flexibly integrate sensorimotor control with analysis of environmental statistics to manage competing visual demands.Significance Statement In everyday behaviour, we often perform manual tasks while simultaneously monitoring the environment, creating competition for gaze. How the brain resolves this competition remains poorly understood. Using a novel paradigm combining an object manipulation task with visual event monitoring, we show that participants integrate knowledge of sensorimotor demands and temporal regularities in the monitoring task to manage gaze. Specifically, we found that participants preferentially allocated gaze to the action task when it is most critical for sensorimotor control and when the likelihood of a visual event was low. Additionally, participants adjusted their hand movement timing based on event statistics to reduce gaze competition. These findings reveal how the brain dynamically allocates gaze resources across competing sensorimotor and visual task demands.

As video transmission increasingly serves machine vision systems (MVS) instead of human vision systems (HVS), video coding for machines (VCM) has become a critical research topic. Existing VCM methods often bind codecs to specific downstream models, requiring retraining or supervised data, thus limiting generalization in multi-task scenarios. Recently, unified VCM frameworks have employed visual backbones (VB) and visual foundation models (VFM) to support multiple video understanding tasks with a single codec. They mainly utilize VB/VFM to maintain semantic consistency or suppress non-semantic information, but seldom explore how to directly link video coding with understanding under VB/VFM guidance. Hence, we propose a Symmetric Entropy-Constrained Video Coding framework for Machines (SEC-VCM). It establishes a symmetric alignment between the video codec and VB, allowing the codec to leverage VB's representation capabilities to preserve semantics and discard MVS-irrelevant information. Specifically, a bi-directional entropy-constraint (BiEC) mechanism ensures symmetry between the process of video decoding and VB encoding by suppressing conditional entropy. This helps the codec to explicitly handle semantic information beneficial to MVS while squeezing useless information. Furthermore, a semantic-pixel dual-path fusion (SPDF) module injects pixel-level priors into the final reconstruction. Through semantic-pixel fusion, it suppresses artifacts harmful to MVS and improves machine-oriented reconstruction quality. Experimental results on classical video understanding tasks and MLLM-based tasks show state-of-the-art (SOTA) rate-task performance. It achieves significant bitrate savings over H.266/VVC reference software VTM on video instance segmentation (avg. 37.4%), video object segmentation (29.8%), object detection (46.2%), multiple object tracking (44.9%), and MLLM-based video grounding (avg. 97.6%). We will release our code soon.

An urgent need exists to elevate the agenda of family ethics, rooted in family nursing's moral commitment to the well-being of families and society. This paper highlights the practice, research, and educational implications for this agenda. Grounded in relational ethics, this vision calls for a curriculum and practice landscape that reflects the lived realities of diverse families and communities. Clinically, it recommends the creation of interdisciplinary care protocols that embed structured family ethics conversations, drawing on an understanding of family systems. In research, it advocates for expanding studies that examine how familial and cultural values shape moral distress and care outcomes, and for broadening traditional bioethical inquiries to include family-centered perspectives. Educationally, it urges the integration of real-world scenarios, cultural analysis, and emotional intelligence training into classrooms and skills labs. Together, these strategies aim to cultivate ethically grounded, socially responsive nurses equipped to lead transformative changes in family health care.

In nature, fruit flies Drosophila have evolved a simple but superb vision system characterised by a three-level synergy: natural compound eyes for panoramic perception, head muscles for continuous tracking in dim conditions, and neural circuits for dynamic scenes. This vision system serves as an ideal model for biomimicry, yet achieving true replication remains challenging, despite significant progress in artificial compound eyes recently. In this study, we present a flexible artificial compound eye camera that adopts such a three-level synergy. An artificial compound eye is constructed by plastic optical fibres and curved microlens arrays for real-time panoramic imaging. Two tethers simulate head muscle movements, achieving a 270&#xb0; field of view for continuous tracking of weak signals. An artificial intelligence algorithm mimics neural processing, enabling the reconstruction of interactive mixed-reality scenes at rates up to 7000 fps. This unique integration of panoramic sensing, active tracking, and neural-like processing establishes a framework for vision-based metaverse applications and bioinspired wearable technologies.

Routine financial activities are now conducted primarily through digital channels. Many such systems remain inaccessible to more than 2.2 billion people globally living with vision impairment, limiting independent financial management. Constrained access can create financial strain and social disadvantage, reducing access to health-enabling resources, and contributing to avoidable health inequities. This scoping review maps evidence on the accessibility of digital financial services for individuals with visual impairment (VI) as a digital determinant of health. We synthesized barriers and facilitators, characterized study designs, settings, and populations, and identified evidence gaps to inform inclusive design, digital health research priorities, and policy. A scoping review was conducted using the Joanna Briggs Institute framework and reported in line with PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines. Eight databases (PubMed, MEDLINE, CINAHL, Scopus, Web of Science, Business Source Complete, ProQuest, and IEEE Xplore) were searched for peer-reviewed papers in English published between 1995 and 2026. Searches featured controlled vocabulary and free-text terms structured in 3 conceptual blocks (VI, digital financial services, and accessibility or usability). A random sample of 20% of titles, abstracts, full texts, and included studies was independently screened or charted by 2 reviewers to calibrate decisions; the remainder were screened and charted by a single reviewer. Data were charted using a standardized extraction form, and results were synthesized descriptively and thematically. Twenty-three studies met the inclusion criteria. Studies were conducted across 12 countries, with the largest number from India (n=7), Indonesia (n=2), Thailand (n=2), and the United States (n=2). Study designs included qualitative studies (n=6), mixed methods studies (n=1), cross-sectional studies (n=4), nonrandomized experimental studies (n=2), and technical or design-focused evaluations (n=6). One study was a large population survey (n=19,136), and the remaining studies with human participants had sample sizes ranging from 4 to 36 participants. Accessibility barriers were reported across all platform types, with authentication-related barriers described in 18 studies and screen reader incompatibility in 17 studies. Reported barriers included reliance on sighted assistance for tasks such as login, verification, and payments, compromising privacy and independence. Facilitators included assistive technology support, logical navigation order, nonvisual feedback mechanisms, and accessible authentication alternatives. Evidence mapping revealed recurrent barrier patterns across Android, iOS, and web platforms. No longitudinal or intervention-based evaluations were identified. This review provides a focused synthesis of accessibility evidence at the intersection of digital financial services and VI, a domain addressed by neither prior digital accessibility reviews nor financial inclusion for people with disabilities. Authentication methods, interface labeling, and navigation were identified as persistent cross-platform accessibility barriers. The findings carry implications for financial technology developers, accessibility auditors, and policymakers implementing accessibility legislation and extend the digital determinants of health framework by demonstrating how inaccessible financial technology may compound health inequities.

Limbal stem cell deficiency (LSCD) is a severe ocular condition that impairs vision and causes chronic pain. Current therapeutic approaches, such as stem cell transplantation and biomaterial scaffolds, are under continuous refinement. This study evaluated a novel treatment strategy combining a biocompatible and biodegradable chitosan-gelatin (CS-Gel) electrospun scaffold with a topical limbal stem cell (LSC) spray for treating alkaline-induced LSCD in a rabbit model. The scaffold was fabricated via electrospinning, stabilized with water vapor, and characterized using SEM, ATR-FTIR, contact angle, weight loss, and swelling analyses. Limbal stem cells were isolated from cadaveric corneas, cultured, and prepared as a spray suspension (104 cells/mL). Nine rabbits with induced alkaline burns were divided into three groups: an untreated control, a group treated with the CS-Gel scaffold alone, and a group receiving the CS-Gel scaffold followed by the LSC spray one week later. Clinical and histopathological evaluations assessed epithelialization, stromal integrity, neovascularization, and inflammation. The results demonstrated that the combined CS-Gel + cell spray treatment significantly enhanced corneal regeneration, reduced inflammation, and improved healing outcomes compared to the other groups. This study confirms the potential of the CS-Gel + cell spray as a promising platform for LSCD therapy, warranting further investigation. Clinical Registration: This study is an animal-based research and does not involve human clinical trials.

Endophthalmitis caused by Nocardia brasiliensis is extremely rare and typically affects immunocompromised individuals, frequently leading to severe vision loss due to diagnostic delays. We report a case of N. brasiliensis endophthalmitis in an older man without prior history of systemic immunosuppression but with newly identified diabetes mellitus, characterized by an indolent initial course followed by fulminant progression. A 67-year-old man without known systemic immunosuppression presented with a two-month history of recurrent right-eye pain and redness, followed by rapid vision loss and a hypopyon. Aqueous humor analysis and metagenomic sequencing identified N. brasiliensis. Despite intravitreal amikacin, systemic antimicrobial therapy, and subsequent pars plana vitrectomy with silicone oil tamponade, intraocular inflammation advanced, resulting in worsening corneal opacification, irreversible structural damage, and a final best-corrected visual acuity of light perception. N. brasiliensis endophthalmitis may progress rapidly and result in severe, irreversible ocular damage, even in patients without overt systemic immunodeficiency. Early microbiologic identification and prompt, targeted antimicrobial therapy combined with timely surgical intervention are critical, although visual outcomes may remain poor in advanced cases.

The classification of fashion images is an essential task in the e-commerce sector, where accurate categorization improves user experience and refines product discovery. Convolutional Neural Networks (CNNs) and Transformers have demonstrated strong performance in image classification tasks due to their ability to learn complex visual features. However, deep variants of these architectures, such as VGG-19, ResNet-50, Vision Transformer (ViT), and Swin Transformer, contain tens of millions of parameters, requiring high memory and powerful GPUs for training, which makes them less suitable for low-resource and edge device environments. To address these limitations, this research proposes a lightweight hybrid architecture, TinyCNN with Linear Self-Attention (LSA), optimized for resource-constrained settings. The proposed model contains fewer than half a million parameters and is trainable on a CPU, achieving a classification accuracy of 91.47% on the Fashion-MNIST dataset. In addition, multiple Explainable Artificial Intelligence (XAI) techniques are implemented, including Self-Attention visualization, Multi-Head Attention, Attention Flow, Attention Rollout, Fixed query position attention maps, Integrated Gradients, LIME, and SHAP, to provide visual interpretability of the model's predictions and enhance transparency in its decision-making process.

Research suggests that reasoning about occluded objects, reality monitoring, and especially planning develop relatively late in children, and are capacities not universally shared with other land animals. F&M's proposal needs to more clearly delineate their use of these concepts to avoid misunderstandings about what they see as essential for the evolution of conscious vision.

Hospital falls can be reduced through patient and staff education, yet limited evidence exists about how staff can systematically implement patient falls prevention education. Planning implementation with staff may enhance their acceptance, engagement, and delivery of falls education to hospital patients. The objective of the study was to design an implementation plan with hospital staff to guide the successful delivery of patient falls education. Three participatory workshops using a world caf&#xe9; methodology were conducted in 1 Western Australian and 2 Victorian hospitals. Participants were presented with information about a patient falls education program called "Safe Recovery" and discussed program implementation strategies. Conversation topics were staff education and training needs, ward support, and organizational requirements. Table discussions were captured on paper and analyzed iteratively at the forum. Subsequently, workshop field notes were analyzed using inductive content analysis. Sixty-two hospital staff (n&#x2009;=&#x2009;42 nurses, n&#x2009;=&#x2009;12 allied health, n&#x2009;=&#x2009;8 other) participated in the workshops. Participants considered the implementation process would be enabled at: (1) individual level, by providing accessible and flexible training to optimize staff engagement; (2) ward level, by establishing clear implementation protocols, engaging and supporting team leaders, and (3) ensuring clear communication between staff, patients, and families; and (4) organizational level, by leadership supporting sustained implementation. Group consensus was that it was important to have a single, agreed vision to implement the Safe Recovery Program. Staff engagement facilitated the development of a shared vision and structured plan to implement a patient falls prevention education program on hospital wards.

Older migrants from Sub-Saharan Africa (SSA) living in high-income countries often face cumulative life-course disadvantages that increase their risk of frailty and related adverse health outcomes. Despite this elevated risk, little is known about the specific health deficits contributing to frailty in this population or how these deficits interact to shape frailty severity. This study aimed to validate the factor structure of the Frailty Instrument for Sub-Saharan Africa (FiSSA) among older SSA migrants in Australia and assess the relative contribution of key health domains to frailty severity. We conducted a cross-sectional survey among 205 SSA migrants aged &#x2265;&#x2009;50&#x2009;years living in Australia. Frailty was assessed using 15 deficits grouped into four domains: Physical, Psychological, Vision, and Socio-emotional. We used confirmatory factor analysis (CFA) to evaluate the latent structure of FiSSA, item response analysis to examine deficit performance across the frailty continuum, and relative weight analysis (RWA) with 10&#x2009;000 bootstrap replications to estimate domain contributions to frailty severity. The CFA supported the four-factor model with acceptable fit (CFI&#x2009;=&#x2009;0.94, TLI&#x2009;=&#x2009;0.92, CMIN/DF&#x2009;=&#x2009;2.91) and significant item loadings. Expected item scores demonstrated monotonic performance across the frailty continuum, with physical items showing the highest sensitivity. RWA indicated that the Physical domain contributed most to frailty variance (32.7%), followed by Socio-emotional (25.5%), Psychological (21.8%), and Vision (20.1%). FiSSA demonstrated strong construct validity among older SSA migrants in Australia, supporting its use as a culturally grounded multidimensional screening tool for early detection and targeted interventions in primary care and community screening contexts.

The robust Dutch rose, also known as the Rosa hybrida is distinguished by its vibrant colors, superior product quality, and extended vase life. These rose varieties, originating from Netherlands, have proven highly successful in Indian agricultural conditions and the international export industry. The dataset consists of a total of 1,995 high resolution petal image collected during this research, encompassing petal color categories, such as red, yellow, white, pink, purple, orange, bi-color, and multi-color, as well as health statuses including fresh, dry, and diseased petals. The primary purpose of this dataset is to support machine learning activities in agriculture and specifically for tasks such as automatic petal health evaluation and rose variety categorization. Although the rose flower is scientifically rich and has a wide range of industrial uses, it has not been given much attention in machine learning, especially when compared to other plant-based datasets. This study adds to the accuracy of quality assessment through the use of modern computer vision and machine learning methods, thus helping the agriculture sector, rose-based edible product making, and flavor development industries.

Identifying cell types present in lung biopsies can provide critical information on pathological processes, tissue organization, and organ health, and are valuable in both clinical and research settings. Knowledge of cell types, their distributions, and spatial relationships can assist pathologists and researchers in diagnosis, prognosis, and mechanistic investigations. Multiplex imaging technologies such as Phenocycler&#x2122;, based on co-detection by indexing (CODEX) can provide whole-slide spatial maps of protein expression and can accurately identify cell types; however, CODEX shares limitations including high costs of antibodies and reagents, as well as labor-intensive conjugation and validation steps. In contrast, hematoxylin and eosin (H&E) staining is inexpensive and routinely available, offering the potential for scalable cell-type mapping if robust prediction can be achieved directly from histology. In this work, we develop a deep learning pipeline to automatically detect cell types in H&E-stained lung tissue sections by leveraging ground-truth annotations from paired CODEX images. The dataset comprises over 2.3 million labeled cells, segmented from lung tissue sections obtained from multiple donors, and grouped into five broad classes: epithelial, immune, endothelial, stromal, and contractile. We train a DeepLabV3+ semantic segmentation model with ResNet backbone on patches from these annotated slides and reserve a subset for testing. Our model achieved 51.3% balanced accuracy across five classes, which is ~2.5x the random-guessing baseline of 20%. This framework demonstrates the feasibility of approximating multiplexed imaging-based cell type maps from routine histology.

Current olfactory tests for untrained dogs have limitations due to use of a single odorant, requirement for ambulation, and potential for vision interference. We hypothesized that our new dogSIT odorant test battery measures would be reliable, repeatable, and age-associated in untrained companion dogs. Gently restrained dogs were presented with 18 odorants (in 9 pairs, one pair was subsequently excluded from analysis); odorants were contained in identical stainless-steel tea-balls. Odorants were presented for 30s in a habituation-dishabituation paradigm (4 consecutive presentations of the same odorant followed by a different odorant). Interaction times were measured post-hoc using masked video analysis; outcome measures represented novel interactions, habituation, and dishabituation. We assessed inter-rater reliability and test-retest repeatability (n&#x2009;=&#x2009;25); and analyzed age, sex, and bodyweight covariates (n&#x2009;=&#x2009;65). Outcome measures had significant inter-rater and test-retest repeatability. There was heterogeneity in odorant preference between dogs. Age, and to a lesser extent bodyweight and sex, were associated with outcome measures relating to novel odorant interaction and dishabituation. Our dogSIT battery is reliable, repeatable, and identified reduced odorant interaction behaviors in older dogs. Our findings have implications for research on food motivation and scent discrimination in aging dogs.