Search Results (66)

Cattle face recognition is a feasible way for identification of cattle in information management of large farms or identity verification in commercial insurance for farms. Recent cattle face recognition approaches, based on supervised learning, heavily depend on annotation which is both labor-intensive and time-consuming. Unsupervised learning for cattle face recognition aims at learning discriminative representations for cattle retrieval from unlabeled data. However, the inherent noise in pseudo-labels significantly hinders the performance. Thus, we propose an unsupervised learning framework with part-attention-based pseudo-label refinement reciprocal compact loss (USL-PARC) to enhance the reliability of the pseudo-label by the fine-grained local context derived via attention mechanism, while obtaining separable and discriminative features by contrastive learning with the compact loss. Firstly, we propose a part-attention-based pseudo-label refinement framework to refine the pseudo-labels of global features by dynamically supplementing local fine-grained information, thereby mitigating the effects of pseudo-label noise. Secondly, ResNet-Sim network, augmented with the SimAM attention mechanism, is constructed to strengthen the ability of capturing more informative localized supplementary information. Finally, we raise compact loss to increase the tightness of the clustering of feature points from the same identity in the feature space. It is encouraging to find that USL-PARC achieves 97.4% accuracy, outperforming the state-of-the-art unsupervised learning models on our CattleFace2025 dataset. These results demonstrate the effectiveness of our proposed USL-PARC on mitigating the impact of pseudo-label noise and enhancing the learning ability of separable and discriminative features. Full article

Deep neural networks (DNNs) have shown strong performance in synthetic aperture radar (SAR) image classification. However, their “black-box” nature limits interpretability and poses challenges for robustness, which is critical for sensitive applications such as disaster assessment, environmental monitoring, and agricultural insurance. This study systematically evaluates the adversarial robustness of five representative DNNs (VGG11/16, ResNet18/101, and A-ConvNet) under a variety of attack and defense settings. Using eXplainable AI (XAI) techniques and attribution-based visualizations, we analyze how adversarial perturbations and adversarial training affect model behavior and decision logic. Our results reveal significant robustness differences across architectures, highlight interpretability limitations, and suggest practical guidelines for building more robust SAR classification systems. We also discuss challenges associated with large-scale, multi-class land use and land cover (LULC) classification under adversarial conditions. Full article

Worker safety has been relatively overlooked in the rapidly growing aquaculture industry. To address this gap, industrial accident compensation insurance data—mainly from floating cage and seaweed farming—were analyzed to quantify accident types and frequencies, with a focus on human elements as root causes. Basic causes were selected based on IMO Resolution A/Res.884 and assessed through a worker awareness survey. Fault Tree Analysis (FTA), a Formal Safety Assessment technique, was applied to evaluate risks associated with these causes. The analysis identified organization at the farm site (23.3%), facility and equipment factors (22.8%), and people factors (21.4%) as the primary causes. Among secondary causes, personal negligence (13.2%), aging gear and poor maintenance (11.4%), and insufficient risk training (10.4%) were the most significant. Selective removal of these causes reduced the probability of human element-related accidents from 64.6% to 48.6%. While limited in scope to Korean data and self-reported surveys, the study demonstrates the value of combining quantitative data with worker perspectives. It provides foundational data for developing tailored safety strategies and institutional improvements—such as standardized procedures, multilingual education, and inclusive risk management—for sustainable safety in aquaculture. Full article

Aligned with the worldwide shift towards promotional and preventative health care, the South African government has introduced a re-engineered primary health care plan, facilitated through a National Health Insurance (NHI) platform. Community health workers (CHWs) are key role-players in most universal health care systems. This article shares insider perspectives from 31 CHWs in one of the South African NHI pilot districts. Based on their perspectives, the authors share recommendations to strengthen the NHI plan. Recommendations comprise of the inclusion of a dedicated CHW team leader and reporting nurse, ongoing CHW education and training with an accompanying portfolio of evidence, and awareness campaigns on the role of CHWs within the South African re-engineered primary health care plan and NHI platform. Full article

Background: The implementation of radiological artificial intelligence (AI) solutions remains challenging due to limitations in existing testing methodologies. This study assesses the efficacy of a comprehensive methodology for performance testing and monitoring of commercial-grade mammographic AI models. Methods: We utilized a combination of retrospective and prospective multicenter approaches to evaluate a neural network based on the Faster R-CNN architecture with a ResNet-50 backbone, trained on a dataset of 3641 mammograms. The methodology encompassed functional and calibration testing, coupled with routine technical and clinical monitoring. Feedback from testers and radiologists was relayed to the developers, who made updates to the AI model. The test dataset comprised 112 medical organizations, representing 10 manufacturers of mammography equipment and encompassing 593,365 studies. The evaluation metrics included the area under the curve (AUC), accuracy, sensitivity, specificity, technical defects, and clinical assessment scores. Results: The results demonstrated significant enhancement in the AI model’s performance through collaborative efforts among developers, testers, and radiologists. Notable improvements included functionality, diagnostic accuracy, and technical stability. Specifically, the AUC rose by 24.7% (from 0.73 to 0.91), the accuracy improved by 15.6% (from 0.77 to 0.89), sensitivity grew by 37.1% (from 0.62 to 0.85), and specificity increased by 10.7% (from 0.84 to 0.93). The average proportion of technical defects declined from 9.0% to 1.0%, while the clinical assessment score improved from 63.4 to 72.0. Following 2 years and 9 months of testing, the AI solution was integrated into the compulsory health insurance system. Conclusions: The multi-stage, lifecycle-based testing methodology demonstrated substantial potential in software enhancement and integration into clinical practice. Key elements of this methodology include robust functional and diagnostic requirements, continuous testing and updates, systematic feedback collection from testers and radiologists, and prospective monitoring. Full article

The global evolution of pathogens causing early-onset sepsis (EOS), a critical condition in preterm infants, necessitates a re-evaluation of risk factors to develop updated prevention and treatment strategies. This nationwide case–control study in Taiwan analyzed data from the National Health Insurance Research Database, Birth Reporting Database, and Maternal and Child Health Database from 2010 to 2019. The study included 176,681 mother–child pairs with preterm births. We identified 2942 clinical EOS cases from 5535 diagnosed sepsis cases, excluding unlikely cases. A control group of 14,710 preterm neonates without EOS was selected at a 1:5 ratio. Clinical EOS increased since 2017. Adjusted logistic regression identified significant EOS risk factors in preterm infants, including maternal fever, chorioamnionitis, maternal diabetes mellitus, maternal antibiotic usage, very preterm birth, birth weight (all with p < 0.001), maternal pneumonia (p = 0.002), and maternal CS (p = 0.004). Effective treatment of maternal conditions like diabetes, fever, and infections is essential to prevent EOS in preterm infants. Key measures include reducing unnecessary antibiotics or steroids, minimizing unnecessary cesarean sections, avoiding premature or prolonged rupture of membranes (PPROM), and increasing gestational age and neonatal birth weight. High-risk preterm neonates should be closely monitored for EOS and considered for antibiotics when warranted. Full article

Maize lodging is a prevalent stress that can significantly diminish corn yield and quality. Unmanned aerial vehicles (UAVs) remote sensing is a practical means to quickly obtain lodging information at field scale, such as area, severity, and distribution. However, existing studies primarily use machine learning (ML) methods to qualitatively analyze maize lodging (lodging and non-lodging) or estimate the maize lodging percentage, while there is less research using deep learning (DL) to quantitatively estimate maize lodging parameters (type, severity, and direction). This study aims to introduce advanced DL algorithms into the maize lodging classification task using UAV-multispectral images and investigate the advantages of DL compared with traditional ML methods. This study collected a UAV-multispectral dataset containing non-lodging maize and lodging maize with different lodging types, severities, and directions. Additionally, 22 vegetation indices (VIs) were extracted from multispectral data, followed by spatial aggregation and image cropping. Five ML classifiers and three DL models were trained to classify the maize lodging parameters. Finally, we compared the performance of ML and DL models in evaluating maize lodging parameters. The results indicate that the Random Forest (RF) model outperforms the other four ML algorithms, achieving an overall accuracy (OA) of 89.29% and a Kappa coefficient of 0.8852. However, the maize lodging classification performance of DL models is significantly better than that of ML methods. Specifically, Swin-T performs better than ResNet-50 and ConvNeXt-T, with an OA reaching 96.02% and a Kappa coefficient of 0.9574. This can be attributed to the fact that Swin-T can more effectively extract detailed information that accurately characterizes maize lodging traits from UAV-multispectral data. This study demonstrates that combining DL with UAV-multispectral data enables a more comprehensive understanding of maize lodging type, severity, and direction, which is essential for post-disaster rescue operations and agricultural insurance claims. Full article

Despite the increasing efficacy of modern medicine in diagnosing and treating cancer, survivors often face discrimination in employment, economics, insurance, and society. Law no. 193/2023, also known as the “Oncological Oblivion Law”, aims to provide an initial legislative response to discrimination against cancer survivors in Italy. After defining oncological oblivion in Article 1, the Law provides, in Articles 2, 3, and 4, directives to prevent discrimination against cancer survivors in the area of access to banking and insurance services, adoption procedures and access to or retention in employment. The aim of this work is to illustrate the content and the critical aspects of the recent Law 193/2023 in the landscape of European directives. The legislative process at the Chamber of Deputies and the Senate of the Italian Republic has been retraced through the consultation of preparatory works and bills registered on institutional databases. Law 193/2023 represents the first initiative in Italy aimed at the recognition of the right to oncological oblivion, not only in access to banking and insurance services as in other countries, but also in adoption, employment, and re-employment. Our opinion piece highlights the need for further clarification and expansion to prevent discrimination and protect the social–work–relational rights of people who have been affected by oncological diseases. Full article

Climate change has brought several natural disasters to South Africa in the form of floods, heat waves, and droughts. Neighbouring countries are also experiencing tropical cyclones, almost on a yearly basis. The insurance sector is faced with an increased level of climate change risk with individuals, corporates, and even the government approaching it for financial cover. However, with an increased level of competition in the insurance sector, (re)insurers must engage in massive product research and development. Therefore, this paper looks at the possibility of the insurance industry developing new products in the form of catastrophe resilience bonds (CAT R Bonds). A qualitative approach is used following content analysis of (re)insurers’ product development policies, marketing documents, company reports, and risk management reports as well as the Conference of Parties 27 and 28 resolution papers. The findings reveal that (re)insurers’ underwriting capacity, reinsurance protection, and innovative and creative product development increase because of CAT R Bonds. CAT R Bonds enhance the interaction between the capital market and money market, thereby giving speculative investors another investment option. Increased investment into new product development such as CAT R Bonds must continue in South Africa in pursuit of climate change resilience goals. Full article

A mediation effect refers to the effect transmitted by a mediator intervening in the relationship between an exposure variable and a response variable. Mediation analysis is widely used to identify significant mediators and to make inferences on their effects. The Bayesian method allows researchers to incorporate prior information from previous knowledge into the analysis, deal with the hierarchical structure of variables, and estimate the quantities of interest from the posterior distributions. This paper proposes three Bayesian mediation analysis methods to make inferences on mediation effects. Our proposed methods are the following: (1) the function of coefficients method; (2) the product of partial difference method; and (3) the re-sampling method. We apply these three methods to explore racial disparities in the diagnostic age of breast cancer patients in Louisiana. We found that African American (AA) patients are diagnosed at an average of 4.37 years younger compared with Caucasian (CA) patients (57.40 versus 61.77, $p<$ 0.0001). We also found that the racial disparity can be explained by patients’ insurance (12.90%), marital status (17.17%), cancer stage (3.27%), and residential environmental factors, including the percent of the population under age 18 (3.07%) and the environmental factor of intersection density (9.02%). Full article

This is a study of decision problems under two-dimensional risk. We use an existing index of absolute correlation aversion to conveniently classify bivariate preferences, with respect to attitudes toward this risk. This classification seems to be more important than whether decision makers are correlation-averse or correlation-seeking for the study of insurance demand when a loss has a multidimensional impact. On this note, we also re-examine Mossin’s theorem under bivariate preferences, where full insurance is preferred with a fair premium, while less than full coverage is preferred with a proportional premium loading. Furthermore, based on the comparative statics of this two-dimensional insurance model for changes in correlation aversion, we derive testable implications about the classification of bivariate utility functions. For the particular case when the two-dimensional risk can be interpreted as risk on income and health, we identify the form of separable utility functions depending on health status and income that is consistent with household disability insurance decisions. Full article

While the impact of the pandemic has varied between and within countries, there are few published data on the relationship between social determinants of health (SDoH) and COVID-19 in Africa. This ecological cross-sectional study examines the relationship between COVID-19 risk and SDoH among 28 African countries. Included were countries with a recent demographic and health survey (years 2010 to 2018). The response variables were COVID-19 case rates and death rates (reported as of 15 August 2020); and the covariates comprised eight broad topics common to multiple SDoH frameworks aggregated to the country level: geography (urban residence), wealth index, education, employment, crowding, and access to information. A negative binomial regression was used to assess the association between aspects of SDoH and COVID-19 outcomes. Our analysis indicated that 1 in 4 (25.1%) households in study countries are without safe and clean water and a space for handwashing. The odds of COVID-19 morbidity and deaths were higher in countries with a high proportion of households without access to safe and clean water. Having a high proportional of educated women (1.003: 95% CI, 1.001–1.005) and living in a less crowded home (0.959: 95% CI, 0.920–1.000) were negatively associated with COVID-19 deaths, while being insured and owning a mobile phone predicted illness. Overall, aspects of SDoH contribute either negatively or positively to COVID-19 outcomes. Thus, addressing economic and environmental SDoH is critical for mitigating the spread of COVID-19 and re-emerging diseases on the African continent. Full article

Chronic obstructive pulmonary disease (COPD) is a progressive respiratory condition characterized by persistent inflammation in the airways, resulting in narrowing and obstruction of the air passages. The development of COPD is primarily attributed to long-term exposure to irritants, such as cigarette smoke and environmental pollutants. Among individuals hospitalized for exacerbations of COPD, approximately one in five is readmitted within 30 days of discharge or encounters immediate post-discharge complications, highlighting a lack of adequate preparedness for self-management. To address this inadequate preparedness, transitional care services (TCS) have emerged as a promising approach. Therefore, this study primarily aims to present a detailed protocol for a multi-site, single-blind, randomized, controlled trial (RCT) aimed at enhancing self-management competency and overall quality of life for patients with COPD through the provision of TCS, facilitated by a proficient Clinical Research Coordinator. The RCT intervention commenced in September 2022 and is set to conclude in December 2024, with a total of 362 COPD patients anticipated to be enrolled in the study. The intervention program encompasses various components, including an initial assessment during hospitalization, comprehensive self-management education, facilitation of social welfare connections, post-discharge home visits, and regular telephone monitoring. Furthermore, follow-up evaluations are conducted at both one month and three months after discharge to assess the effectiveness of the intervention in terms of preventing re-hospitalization, reducing acute exacerbations, and enhancing disease awareness among participants. The results of this study are expected to provide a basis for the development of TCS fee payment policies for future health insurance. Full article

The COVID-19 pandemic has created massive challenges for women’s employment. Women’s responsibilities were exacerbated by the closure of schools and child daycare facilities. Investigating the determinants of job losses among women is critical to avoiding dropouts and supporting re-entry into the labor market. This study investigates the factors driving women’s workforce losses during the pandemic in five Arab countries (Egypt, Tunisia, Morocco, Jordan, and Sudan). The current study focuses mainly on how COVID-19-induced responsibilities affected women’s employment during the pandemic. The study depends on the COVID-19 MENA Monitor Household Survey produced by the Economic Research Forum. The factor analysis of mixed data is used to construct the women’s responsibilities index that is made up of 18 variables. The mixed-effect logistic model is used to consider changes in working arrangements across economic activities. The results indicate that women with high family caregiving responsibilities were more likely to lose their jobs. Women working in the government sector and with health insurance were protected from job losses. Telecommuting played a significant role in helping women maintain their jobs. Work arrangements should be improved to consider increased unpaid domestic work. Family-friendly policies must be activated, and childcare leave must be facilitated and funded. The private sector should also be urged to improve workplace flexibility. Full article