Search Results (792)

Background/Objectives: Diabetic retinopathy is a leading cause of vision impairment worldwide, and the development of reliable automated classification systems is crucial for early diagnosis and clinical decision-making. This study presents a comprehensive comparative evaluation of two state-of-the-art deep learning families for the task of classifying diabetic retinopathy using retinal fundus images. Methods: The models were trained and tested in both binary and multi-class settings. The experimental design involved partitioning the data into training (70%), validation (20%), and testing (10%) sets. Model performance was assessed using standard metrics, including precision, sensitivity, specificity, F1-score, and the area under the receiver operating characteristic curve. Results: In binary classification, the ResNeXt101-64x4d model and RegNetY32GT model demonstrated outstanding performance, each achieving high sensitivity and precision. For multi-class classification, ResNeXt101-32x8d exhibited strong performance in early stages, while RegNetY16GT showed better balance across all stages, particularly in advanced diabetic retinopathy cases. To enhance transparency, SHapley Additive exPlanations were employed to visualize the pixel-level contributions for each model’s predictions. Conclusions: The findings suggest that while ResNeXt models are effective in detecting early signs, RegNet models offer more consistent performance in distinguishing between multiple stages of diabetic retinopathy severity. This dual approach combining quantitative evaluation and model interpretability supports the development of more robust and clinically trustworthy decision support systems for diabetic retinopathy screening. Full article

Background: The long-term clinical and laboratory results of a 33-year follow-up of a Greek family with abetalipoproteinemia (ABL) are described. Case Report: The patients (two brothers and their sister, aged 57, 49, and 62 years, respectively) are still alive, being under close surveillance. In two of the three patients, diarrhea appeared in early infancy, while in the third, it appeared during adolescence. CNS symptomatology worsened after the second decade of life. At the same time, night blindness appeared in the advanced stages of the disease, resulting in almost complete loss of vision in one of the male patients and severe impairment in the other. The diagnosis was based on the clinical picture, ophthalmological findings, serum lipid estimations, and presence of peripheral acanthocytosis. All patients exhibited typical serum lipidemic profile, ophthalmological findings, and acanthocytes in the peripheral blood. During the follow-up period, strict dietary modifications were applied, including the substitution of fat with medium-chain triglycerides (MCT oil). After 33 years since the initial diagnosis, all patients are alive without any sign of liver dysfunction despite continuous use of MCT oil. However, symptoms from the central nervous system and vision impairment worsened. Conclusion: The course of these patients suggests that the application of a modified diet, including MCT oil, along with close surveillance, could prolong the survival of patients without significant side effects from the liver. Full article

Background/Objectives: Keratoconus (KC) is a bilateral asymmetric corneal ectasia characterized by progressive corneal thinning, irregular astigmatism, and impaired visual acuity. The National Eye Institute (NEI) developed the Visual Function Questionnaire (VFQ-25) to assess the impact of visual impairment on quality of life. This study aimed to evaluate the effect of penetrating keratoplasty (PKP) on quality of life and visual acuity in KC patients one year postoperatively. Methods: A retrospective study was conducted between January 2018 and December 2022 at the Ophthalmology Department of Saint Barbara Hospital, Trauma Center, Sosnowiec, Poland. A total of 71 patients (86 eyes) diagnosed with KC underwent PKP. The VFQ-25 questionnaire and visual acuity measurements were assessed preoperatively and one year postoperatively. Results: The study cohort included 71 patients (20 females, 28.17%; 51 males, 71.83%). Preoperative visual acuity ranged from less than 0.05 on the Snellen chart to 0.5. Postoperatively, visual acuity improved to a range of 0.1–1.0. A visual acuity of 1.0 was achieved in 21 eyes (24.42%; 5 females, 24%; 16 males, 76%), with a statistically significant improvement (p < 0.01). The mean VFQ-25 composite score increased from 57.96 (±17.58) preoperatively to 81.42 (±14.66) postoperatively (p < 0.001). Domains with the lowest preoperative scores were “role difficulties,” “general vision,” and “mental health,” while “color vision” scored highest. Conclusions: PKP significantly enhances both objective visual acuity and subjective quality of life in KC patients, as reflected in VFQ-25 questionnaire outcomes. Full article

Background: Mobility limitation is a pivotal but under-documented dimension of disability in Saudi Arabia. Leveraging the 2017 National Disability Survey, this cross-sectional study provides a population-wide profile of mobility-related physical difficulty. Objectives: Five research aims were pursued: (1) estimate national prevalence and severity by sex; (2) map regional differentials; (3) examine educational and marital correlates; (4) characterize cause, duration, and familial context among those with multiple limitations; and (5) describe patterns of assistive-aid and social-service use. Methods: Publicly available aggregate data covering 20,408,362 Saudi citizens were cleaned and analyzed across 14 mobility indicators and three baseline files. Prevalence ratios and χ² tests assessed associations. Results: Overall, 1,445,723 Saudis (7.1%) reported at least one functional difficulty; 833,136 (4.1%) had mobility difficulty, of whom 305,867 (36.7%) had mobility-only impairment. Severity was chiefly mild (35% of cases), with moderate (16%) and severe (7%) forms forming a descending pyramid. Prevalence varied more than threefold across the thirteen regions, peaking in Aseer (9.4%) and bottoming in Najran (2.9%). Mobility difficulty clustered among adults with no schooling (36.1%) and widowed status (18.5%), with sharper female disadvantage in both domains (p < 0.001). Among those with additional limitations, chronic disease dominated etiology (56.3%), and 90.1% had lived with disability for ≥25 years; women were overrepresented in the longest-duration band. Aid utilization was led by crutches (47.7%), personal assistance (25.3%), and wheelchairs (22.6%), while 83.8% accessed Ministry rehabilitation services, yet fewer than 4% used home or daycare support. Conclusions: These findings highlight sizeable, regionally concentrated, and gender-patterned mobility burdens, underscoring the need for education-sensitive prevention, chronic-care management, investment in advanced assistive technology, and distributed community services to achieve Vision 2030 inclusion goals. Full article

Stargardt’s disease (STGD1) is an autosomal recessive juvenile macular degeneration caused by mutations in the ABCA4 gene, impairing clearance of toxic retinoid byproducts in the retinal pigment epithelium (RPE). This leads to lipofuscin accumulation, oxidative stress, photoreceptor degeneration, and central vision loss. Over 1200 pathogenic/likely pathogenic ABCA4 variants highlight the genetic heterogeneity of STGD1, which manifests as progressive central vision loss, with phenotype influenced by deep intronic variants, modifier genes, and environmental factors like light exposure. ABCA4 variants also show variable penetrance and geographical prevalence. With no approved treatment, investigational therapies target different aspects of disease pathology. Small-molecule therapies target vitamin A dimerization (e.g., ALK-001), inhibit lipofuscin accumulation (e.g., soraprazan), or modulate the visual cycle (e.g., emixustat hydrochloride). Gene therapy trials explore ABCA4 supplementation including strategies like RNA exon editing (ACDN-01) and bioengineered ambient light-activated OPSIN. RORA gene therapy (Phase 2/3) addresses oxidative stress, inflammation, lipid metabolism, and complement system dysregulation. Trials like DRAGON (Phase 3, tinlarebant), STARLIGHT (phase 2, bioengineered OPSIN) show promise, but optimizing efficacy remains challenging. With the key problem of establishing genotype–phenotype correlations, the future of STGD1 therapy may rely on approaches targeting oxidative stress, lipid metabolism, inflammation, complement regulation, and genetic repair. Full article

Background/Objectives: Diabetic retinopathy (DR) is a progressive microvascular complication of diabetes mellitus and a leading cause of vision impairment worldwide. Early detection and timely management are critical in preventing vision loss, yet current screening programs face challenges, including limited specialist availability and variability in diagnoses, particularly in underserved areas. This literature review explores the evolving role of artificial intelligence (AI) in enhancing the diagnosis, screening, and management of diabetic retinopathy. It examines AI’s potential to improve diagnostic accuracy, accessibility, and patient outcomes through advanced machine-learning and deep-learning algorithms. Methods: We conducted a non-systematic review of the published literature to explore advancements in the diagnostics of diabetic retinopathy. Relevant articles were identified by searching the PubMed and Google Scholar databases. Studies focusing on the application of artificial intelligence in screening, diagnosis, and improving healthcare accessibility for diabetic retinopathy were included. Key information was extracted and synthesized to provide an overview of recent progress and clinical implications. Conclusions: Artificial intelligence holds transformative potential in diabetic retinopathy care by enabling earlier detection, improving screening coverage, and supporting individualized disease management. Continued research and ethical deployment will be essential to maximize AI’s benefits and address challenges in real-world applications, ultimately improving global vision health outcomes. Full article

Background/Objectives: Senior-Loken syndrome (SLS) is a rare autosomal recessive renal–retinal disease caused by mutations in 10 genes. This study aimed to review the ophthalmic findings, renal function, and genotypes of Korean SLS cases. Methods: We retrospectively reviewed 17 genetically confirmed SLS patients in Korea, including 9 newly identified cases and 8 previously reported. Comprehensive ophthalmologic evaluations and renal assessments were conducted. Genetic testing was performed using whole-genome sequencing (WGS), whole-exome sequencing (WES), or Sanger sequencing. Results: Among the 17 patients, patients with NPHP1 mutations were most common (35.3%), followed by those with NPHP4 (29.4%), IQCB1 (NPHP5, 29.4%), and SDCCAG8 (NPHP10, 5.9%) mutations. Patients with NPHP1 mutations showed retinitis pigmentosa (RP) sine pigmento and preserved central vision independent of renal deterioration. Patients with NPHP4 mutations showed early renal dysfunction. Two patients aged under 20 maintained relatively good visual function, but older individuals progressed to severe retinopathy. Patients with IQCB1 mutations were generally prone to early and severe retinal degeneration, typically manifesting as Leber congenital amaurosis (LCA) (three patients), while two patients exhibited milder RP sine pigmento with preserved central vision. Notably, two out of five (40.0%) maintained normal renal function at the time of diagnosis, and both had large deletions in IQCB1. The patient with SDCCAG8 mutation exhibited both end-stage renal disease and congenital blindness due to LCA. Wide-field fundus autofluorescence (AF) revealed perifoveal and peripapillary hypoAF with a perifoveal hyperAF in younger patients across genotypes. Patients under 20 years old showed relatively preserved central vision, regardless of the underlying genetic mutation. Conclusions: The clinical manifestation of renal and ocular impairment demonstrated heterogeneity among Korean SLS patients according to causative genes, and the severity of renal dysfunction and visual decline was not correlated. Therefore, simultaneous comprehensive evaluations of both renal and ocular function should be performed at the initial diagnosis to guide timely intervention and optimize long-term outcomes. Full article

Background: Arachnoid cysts are typically benign and asymptomatic, but large cysts can exert a mass effect on adjacent neural structures. Based on the available literature, no cases of cortical visual impairment (CVI) in an adolescent caused by posterior fossa arachnoid cysts have been reported. Case presentation: We report the case of a previously healthy 16-year-old girl who presented with sudden and rapidly progressive bilateral visual loss due to a large retrocerebellar arachnoid cyst. She reported blurred vision, tunnel vision-like, and decreased visual acuity. Although neuro-ophthalmologic and imaging workup revealed no damage to the anterior visual pathways, she exhibited progressive visual decline. Functional tests confirmed bilateral cortical visual impairment: pattern-reversal visual evoked potentials (VEPs) showed preserved and symmetric P100 latencies and amplitudes, while automated perimetry revealed bilateral concentric visual field constriction with preserved central islands. Following cystoperitoneal drainage, her vision rapidly and completely recovered. Conclusions: To the best of our knowledge, this is the first reported case of reversible CVI in an adolescent caused by a posterior fossa arachnoid cyst without intracranial pressure (ICP) elevation or optic nerve involvement, and with tunnel vision-like. Our findings emphasize the role of posterior fossa lesions in visual dysfunction and highlight the potential reversibility of cortical visual loss when timely decompression is achieved. This case underscores the importance of including posterior fossa lesions in the differential diagnosis of unexplained bilateral visual loss, even in the absence of elevated intracranial pressure or anterior visual pathway involvement. Full article

Blindness and low vision are increasing serious public health issues that affect a significant percentage of the population worldwide. Vision plays a crucial role in spatial navigation and daily activities. Its reduction or loss creates numerous challenges for an individual. Assistive technology can enhance mobility and navigation in outdoor environments. In the field of orientation and mobility training, technologies with haptic interaction can assist individuals with visual impairments in learning how to navigate safely and effectively using the sense of touch. This paper presents a virtual reality platform designed to support the development of navigation techniques within a safe yet realistic environment, expanding upon existing research in the field. Following extensive optimization, we present a visual representation that accurately simulates various 3D tile textures using graphics replicating real tactile surfaces. We conducted a user interaction study in a virtual environment consisting of 3D navigation tiles enhanced with tactile textures, placed appropriately for a real-world scenario, to assess user performance and experience. This study also assess the usability and user experience of the platform. We hope that the findings will contribute to the development of new universal navigation techniques for people with visual impairments. Full article

Human pose estimation is an important downstream task in computer vision, with significant applications in action recognition and virtual reality. However, data collected in a decentralized manner often exhibit non-independent and identically distributed (non-IID) characteristics, and traditional federated learning aggregation strategies can lead to gradient conflicts that impair model convergence and accuracy. To address this, we propose the Federated Gradient Harmonization aggregation strategy (FedGH), which coordinates update directions by measuring client gradient discrepancies and integrating gradient-projection correction with a parameter-reconstruction mechanism. Experiments conducted on a self-constructed single-arm robotic dataset and the public Max Planck Institute for Informatics (MPII Human Pose Dataset) dataset demonstrate that FedGH achieves average Percentage of Correct Keypoints (PCK) of 47.14% and 66.31% across all keypoints, representing improvements of 1.82 and 0.36 percentage points over the Federated Adaptive Weighting (FedAW) method. On our self-constructed dataset, FedGH attains a PCK of 86.4% for shoulder detection, surpassing other traditional federated learning methods by 20–30%. Moreover, on the self-constructed dataset, FedGH reaches over 98% accuracy in the keypoint heatmap regression model within the first 10 rounds and remains stable between 98% and 100% thereafter. This method effectively mitigates gradient conflicts in non-IID environments, providing a more robust optimization solution for distributed human pose estimation. Full article

Navigational aids play a vital role in enhancing the mobility and independence of blind and visually impaired (VI) individuals. However, existing solutions often present challenges related to discomfort, complexity, and limited ability to provide detailed environmental awareness. To address these limitations, we introduce SnapStick, an innovative assistive technology designed to improve spatial perception and navigation. SnapStick integrates a Bluetooth-enabled smart cane, bone-conduction headphones, and a smartphone application powered by the Florence-2 Vision Language Model (VLM) to deliver real-time object recognition, text reading, bus route detection, and detailed scene descriptions. To assess the system’s effectiveness and user experience, eleven blind participants evaluated SnapStick, and usability was measured using the System Usability Scale (SUS). In addition to the 94% accuracy, the device received an SUS score of 84.7%, indicating high user satisfaction, ease of use, and comfort. Participants reported that SnapStick significantly improved their ability to navigate, recognize objects, identify text, and detect landmarks with greater confidence. The system’s ability to provide accurate and accessible auditory feedback proved essential for real-world applications, making it a practical and user-friendly solution. These findings highlight SnapStick’s potential to serve as an effective assistive device for blind individuals, enhancing autonomy, safety, and navigation capabilities in daily life. Future work will explore further refinements to optimize user experience and adaptability across different environments. Full article

In computer vision, the performance of detectors trained under optimal lighting conditions is significantly impaired when applied to infrared domains due to the scarcity of labeled infrared target domain data and the inherent degradation in infrared image quality. Progress in cross-domain pedestrian re-identification is hindered by the lack of labeled infrared image data. To address the degradation of pedestrian recognition in infrared environments, we propose a framework for zero-shot infrared domain adaptation. This integrated approach is designed to mitigate the challenges of pedestrian recognition in infrared domains while enabling zero-shot domain adaptation. Specifically, an advanced reflectance representation learning module and an exchange–re-decomposition–coherence process are employed to learn illumination invariance and to enhance the model’s effectiveness, respectively. Additionally, the CLIP (Contrastive Language–Image Pretraining) image encoder and DINO (Distillation with No Labels) are fused for feature extraction, improving model performance under infrared conditions and enhancing its generalization capability. To further improve model performance, we introduce the Non-Local Attention (NLA) module, the Instance-based Weighted Part Attention (IWPA) module, and the Multi-head Self-Attention module. The NLA module captures global feature dependencies, particularly long-range feature relationships, effectively mitigating issues such as blurred or missing image information in feature degradation scenarios. The IWPA module focuses on localized regions to enhance model accuracy in complex backgrounds and unevenly lit scenes. Meanwhile, the Multi-head Self-Attention module captures long-range dependencies between cross-modal features, further strengthening environmental understanding and scene modeling. The key innovation of this work lies in the skillful combination and application of existing technologies to new domains, overcoming the challenges posed by vision in infrared environments. Experimental results on the SYSU-MM01 dataset show that, under the single-shot setting, Rank-1 Accuracy (Rank-1) andmean Average Precision (mAP) values of 37.97% and 37.25%, respectively, were achieved, while in the multi-shot setting, values of 34.96% and 34.14% were attained. Full article

With the rapid advancement of urbanization in China, the demand for community-based elderly care facilities (CECFs) has been increasing. One pressing challenge is the question of how to provide CECFs that not only meet the health needs of the elderly but also make efficient use of limited urban land resources. This study addresses this issue by adopting an integrated multi-method research framework that combines multi-objective optimization (MOO) algorithms, Spearman rank correlation analysis, ensemble learning methods (Random Forest combined with SHapley Additive exPlanations (SHAP), where SHAP enhances the interpretability of ensemble models), and Self-Organizing Map (SOM) neural networks. This framework is employed to identify optimal building configurations and to examine how different architectural parameters influence key daylight performance indicators—Useful Daylight Illuminance (UDI) and Daylight Factor (DF). Results indicate that when UDI and DF meet the comfort thresholds for elderly users, the minimum building area can be controlled to as little as 351 m² and can achieve a balance between natural lighting and spatial efficiency. This ensures sufficient indoor daylight while mitigating excessive glare that could impair elderly vision. Significant correlations are observed between spatial form and daylight performance, with factors such as window-to-wall ratio (WWR) and wall thickness (WT) playing crucial roles. Specifically, wall thickness affects indoor daylight distribution by altering window depth and shading. Moreover, the ensemble learning models combined with SHAP analysis uncover nonlinear relationships between various architectural parameters and daylight performance. In addition, a decision support method based on SOM is proposed to replace the subjective decision-making process commonly found in traditional optimization frameworks. This method enables the visualization of a large Pareto solution set in a two-dimensional space, facilitating more informed and rational design decisions. Finally, the findings are translated into a set of practical design strategies for application in real-world projects. Full article

Background/Objectives: As high-acuity vision is limited to a very small visual angle, examination of a scene requires multiple fixations. Simultanagnosia, a disorder wherein elements of a scene can be perceived correctly but cannot be integrated into a coherent whole, has been parsed into dorsal and ventral forms. In ventral simultanagnosia, limited visual integration is possible. This case study was the first to record gaze during the presentation of a series of visual stimuli, which required the processing of local and global elements. We hypothesised that gaze patterns would differ with successful processing and that feature integration could be disrupted by distractors. Methods: The patient received a neuropsychological assessment and underwent CT and MRI. Eye movements were recorded during the following tasks: (1) famous face identification, (2) facial emotion recognition, (3) identification of Ishihara colour plates, and (4) identification of both local and global letters in Navon composite letters, presented either alone or surrounded by filled black circles, which we hypothesised would impair global processing by disrupting fixation. Results: The patients identified no famous faces but scanned them qualitatively normally. The only emotion to be consistently recognised was happiness, whose scanpath differed from the other emotions. She identified none of the Ishihara plates, although her colour vision was normal on the FM-15, even mapping an unseen digit with fixations and tracing it with her finger. For plain Navon figures, she correctly identified 20/20 local and global letters; for the “dotted” figures, she was correct 19/20 times for local letters and 0/20 for global letters (chi-squared NS for local, p < 0.0001, global), with similar fixation of salient elements for both. Conclusions: Contrary to our hypothesis, gaze behaviour was largely independent of the ability to process global stimuli, showing for the first time that normal acquisition of visual information did not ensure its integration into a percept. The core defect lay in processing, not acquisition. In the novel Navon task, adding distractors abolished feature integration without affecting the fixation of the salient elements, confirming for the first time that distractors could disrupt the processing, not the acquisition, of visual information in this disorder. Full article

Glaucoma is an irreversible neurodegenerative disease that affects the optic nerve, leading to partial or complete vision loss. Early and accurate detection is crucial to prevent vision impairment, which necessitates the development of highly precise diagnostic tools. Deep learning (DL) has emerged as a promising approach for glaucoma diagnosis, where the model is trained on datasets of fundus images. To improve the detection accuracy, we propose a hybrid model for glaucoma detection that combines multiple DL models with two fine-tuning strategies and uses a majority voting scheme to determine the final prediction. In experiments, the hybrid model achieved a detection accuracy of 96.55%, a sensitivity of 98.84%, and a specificity of 94.32%. Integrating datasets was found to improve the performance compared to using them separately even with transfer learning. When compared to individual DL models, the hybrid model achieved a 20.69% improvement in accuracy compared to the best model when applied to a single dataset, a 13.22% improvement when applied with transfer learning across all datasets, and a 1.72% improvement when applied to all datasets. These results demonstrate the potential of hybrid DL models to detect glaucoma more accurately than individual models. Full article