Browse Articles
Discover research articles across all indexed journals
Correction: Assessing the risk of bias of clinical trials with large language models and ROBUST-RCT: a feasibility study
Typographic self-portrait generation in an interactive museum installation
Abstract This article presents Photomaton , an interactive museum installation that generates typographic self-portraits of visitors using computational design methods. Operating in a photo booth configuration, the system captures a photograph of the user and renders a portrait composed exclusively of glyphs, combined with a user-selected literary text. We describe the design, development, and evaluation of Photomaton , including three generative approaches for typographic portrait creation based on text, words, and glyphs, as well as a template-based method for automatic typesetting and layout generation. A prototype was evaluated with 46 participants, demonstrating efficient portrait generation, high recognisability of user facial features and strong user satisfaction. The installation has been successfully deployed as part of the permanent Torre Literária exhibition, dedicated to Lusophone literature and inaugurated in Vila Nova de Famalicão, Portugal, in late 2020. The contributions of this work include the development of an interactive system for real-time typographic self-portrait generation; the design and evaluation of multiple generative approaches that explore the balance between textual readability and image recognisability; the integration of automatic typesetting and template-based layout generation within a fully functional interactive installation; and the demonstration of its applicability in real-world museum contexts, including support for print-on-demand customised artefacts for visitors. Additionally, this work provides a documented case study demonstrating the deployment of computational design methods in a public museum setting.
A cloud-based two-layer text classification framework for mental health screening with sarcasm and emoji-aware sentiment analysis
Abstract The increasing use of digital communication platforms has led individuals to express emotions and mental health concerns through text containing implicit emotional cues, informal language, and non-standard expressions. Traditional sentiment analysis systems often struggle to capture these contextual nuances, limiting their effectiveness in mental health-related text analysis . To address this challenge, this study proposes a two-layer framework that combines Azure Sentiment Analysis and Azure Custom Text Classification for sentiment and mental health-related text categorization. In the first layer, user-generated text is classified into positive, neutral, or negative sentiment categories using Azure Sentiment Analysis. Text identified as negative is subsequently analysed using Azure Custom Text Classification to categorize content into predefined mental health-related classes, including Anxiety, Depression, PTSD, Social Anxiety Disorder, and Suicidal Ideation and Behaviour. The proposed framework aims to provide a structured approach for identifying linguistic patterns associated with mental health-related discussions and supporting mental health screening and triage applications. Experimental evaluation using an 80% training and 20% testing split achieved an overall Precision, Recall, and F1-score of 96.97%. Class-level evaluation demonstrated strong performance across multiple categories, with F1-scores ranging from 0.94 to 1.000. The findings indicate that the proposed architecture can effectively classify mental health-related textual content within the evaluated dataset while providing a scalable framework for automated sentiment and text classification. The study contributes to the growing field of intelligent emotional computing and highlights the potential of cloud-based natural language processing tools for mental health-related text analytics . The reported results are limited to the evaluated dataset and should be interpreted as a text classification and screening approach rather than a clinical diagnostic system. This manuscript presents the computational component of a broader mixed-methods study registered under CTRI/2024/06/068766, titled “Exploring Mental Health Status in a Selected Population: A Corpus Analysis Combining Forensic Linguistics and Psychology - a Mixed Method Study.” The current work focuses on the development and validation of an AI-based diagnostic tool for mental health assessment using synthetic and anonymized textual data, constituting a secondary objective of the registered protocol. Registry: Clinical Trials Registry- India (CTRI) Trial Registration Number: CTRI/2024/06/068766 Date of Registration: 12.06.2024.
Publisher Correction: The involvement of microglia and the CXCL16-CXCR6 axis in the recruitment of CD8+ T cells to an amyloidogenic mouse brain
IL-17 A inhibition attenuates pulmonary inflammation following blast lung injury
Automated camouflage pattern design based on conditional generative adversarial network and image quilting
Finite element analysis and case series study of anchor loop plate fixation of patellar inferior pole sleeve avulsion fracture
Mitigating request flooding attack in named data networking using federated learning
Abstract Named Data Networking (NDN) represents a paradigm shift toward content-centric architectures but remains critically vulnerable to Interest Flooding Attacks (IFAs), where malicious actors overwhelm router Pending Interest Tables with spurious requests, causing service degradation and denial-of-service. To address the limitations of existing approaches, including high false positives in threshold-based methods and substantial overhead in centralized learning, we propose , a novel federated learning framework for adaptive IFA mitigation. Our solution integrates dynamic Poisson-EMA thresholding for accurate flood detection, entropy-aware federated aggregation to handle non-IID traffic distributions across edge routers, and Byzantine-robust mechanisms with differential privacy guarantees. Comprehensive evaluation on the FIT/IoT-LAB testbed with 100 routers demonstrates exceptional performance: 93.1% F1-score in attack detection, only 5% false positives, 28 ms average end-to-end latency ( $$\Delta t_{\text {e2e}}$$ ), and over 90% legitimate Interest Satisfaction Ratio under sophisticated collusive attacks, while maintaining minimal computational overhead (<9% CPU utilization on ARMv8 routers). significantly improves security performance, offering 35% higher accuracy than static thresholding and 60% lower communication overhead than centralized approaches. While simpler heuristic baselines naturally incur marginally lower computational footprints, our solution delivers the optimal overall operational balance among high precision, low end-to-end latency ( $$\Delta t_{\text {e2e}}$$ ), and resource efficiency in constrained edge computing environments.
Engineering and predicting the environmental impact of H2-O2-enriched liquefied petroleum gas blend for reducing pollutant emissions with high burning potential
Abstract A large fraction of global thermal energy demand will remain combustion-based for decades, making the decarbonization of gaseous fuels a critical challenge. Liquefied petroleum gas (LPG) remains widely used, particularly where rapid electrification is constrained, but its propane–butane composition results in significant carbon emissions. In this study, we demonstrate that controlled reformulation of LPG with hydrogen and stoichiometric oxygen can reduce carbon intensity while preserving combustion performance. LPG-H 2 -O 2 blends were experimentally evaluated through emission measurements (CO 2 , NO, NO x ), burning potential analysis, and a multi-criteria decision matrix. The optimal blend, formulated through LPG replacement with 40% H 2 and O 2 corresponding to 30% of the blend stoichiometric oxygen demand, achieved substantial reductions in CO 2 emissions and LPG consumption while maintaining stable operation.. To assess broader impacts, experimental results were integrated with demographic-based projections of LPG demand across all Brazilian states. Under a full substitution scenario, cumulative avoided emissions could reach ~ 2.9 × 10 8 tons of CO 2 by 2045 without requiring infrastructure changes. These findings highlight hydrogen-assisted LPG reformulation as a scalable decarbonization strategy for combustion-based energy systems.
Deep learning integrated plasmonic electrochemical sensing for fast and accurate pathogen detection
Abstract The development of plasmonic electrochemical biosensors using the new generation of deep learning algorithms is a potent pathway toward the troublesome, immediate and field-mediable diagnostics of the pathogen. This article provides a combination of a MobileNet-Transformer and Gated Recurrent Unit (GRU) deep neural network with a nanostructured plasmonic biosensor designed to sense Escherichia coli , Salmonella typhimurium , and Staphylococcus aureus at an early stage. To augment the charge-transfer kinetics in the biosensor, localized surface plasmon resonance (LSPR) is utilized by use of gold-nanoparticle graphene oxide hybrid nanocomposites which lead to maximized electrochemical responses. The platform has ultra-low E. coli , Salmonella , and S. aureus limits of detection of 0.12 pg/mL, 0.17 pg/mL and 0.21 pg/mL, respectively using 5 μL of sample and a time of assay of less than 10 min. The deep learning pipeline processes raw voltammetric signals automatically with MobileNet-Transformer being helpful to determine the features effectively and GRU to reduce the noise related to time. The system was better than baseline CNN and RNN models, with a classification accuracy 95.6% and area under the curve of 0.986 as well as better precision-recall profiles. The vehicular combinations of plasmonic enhancement and deep learning deposition make it possible to realize real-time on-device decision-making that can be made applicable in food safety checks, environmental or point-of-care diagnoses. This paper illustrates a scalable path to AI-assisted electrochemical biosensing and a similar performance on par with laboratory benchtop systems and that is fully compatible with low-cost diagnostic hardware in a portable format.
Atypical genomic features among Listeria monocytogenes strains isolated from the pig manure management chain
Abstract Listeria monocytogenes is a major foodborne pathogen found in outdoors, wild and domesticated animals, food-processing environments and food products. Genomic studies have mainly focused on strains from clinical or food-associated sources, but other reservoirs have been mostly underexplored. We performed in-depth phylogenetic and pangenome analyses of 371 L. monocytogenes strains isolated along the pig manure management chain, from living pigs to treated manure. Chromosome and plasmid-related virulence and resistance markers were thoroughly examined. Genomic islands along with mobile genetic elements were characterized. Anaerobic digestion and nitrification–denitrification treatments did not significantly impact the clonal diversity. We identified a novel Listeria Genomic Island 3 (LGI3) variant carrying a new internalin gene in 46 strains of Clonal Complex CC77. Chromosomal resistance to tetracycline was found in 7% of strains, linked to one Tn5801 and four distinct Tn916 acquisition events. In total, 100 plasmids were identified and characterized, including several multi–heavy metal resistance plasmids and one associated with multidrug antibiotic resistance. We characterized a large diversity of 932 prophages integrated across 10 chromosomal sites. This study provides innovative genomic insights into L. monocytogenes populations within the pig manure value chain and highlights potential routes of transmission between pig, farm environment and humans.
Policy-driven municipal solid waste network optimization under carbon regulation: a risk-informed MILP framework for a post-conflict recovery city
Effect of tissue preparation and storage on elemental expression in oral tissues using
Abstract Laser-induced breakdown spectroscopy (LIBS) is a recently emerging novel spectroscopic technique for element detection that has been used in various fields, including biomedical applications. In clinical settings, it has been used in detecting elemental composition in samples of various complexities, such as human body fluids, on different hard and soft tissues with no or minimal sample preparation. In this work, we aimed to analyze the effects of tissue preparation and storage on the expression of elements in LIBS. These tissue preparation techniques have been divided into 3 categories, namely, fresh frozen, formalin fixed paraffin embedded tissue blocks (FFPE), and formalin fixed tissue sections (FFS). The expression of elements using LIBS in fresh frozen oral tissues showed better results compared with the FFPE tissue blocks and FFS owing to the interference of paraffin in the latter two, making fresh frozen tissue samples preferable for LIBS analysis. Extensive research based on a large sample size needs to be implemented to standardize the protocol for tissue preparation before using LIBS for elemental analysis.
CT radiomics for survival risk stratification in resectable colorectal liver metastases: a multi-centre study
Structure-based computational screening and molecular dynamics reveal potential inhibitors of Norovirus VP1 and RdRp Proteins: an in-silico study
Abstract Norovirus is a major enteric pathogen with pandemic potential and disproportionately high mortality in low-income countries, particularly among young children. Despite its global health burden, no approved vaccine or specific antiviral therapy is currently available. In this study, we targeted two key viral proteins, viral protein 1 (VP1) and RNA-dependent RNA polymerase (RdRp). The workflow included protein modeling, structural stability assessment, molecular docking, molecular dynamics (MD) simulations, non-covalent interaction (NCI) analysis, protein contact atlas evaluation, and pharmacokinetic (ADME-Tox) profiling. Molecular docking results indicated strong binding affinities of selected phytochemicals—Zingiberol, Cardeonolide, Boeravinone B, β-Elemene, and Fisetin—toward both VP1 and RdRp, with binding energies ranging from − 7.8 to − 9.4 kcal/mol. MD simulations further demonstrated the structural stability of protein–ligand complexes, with stable RMSD values (~ 0.3 nm for RdRp and 0.3–0.5 nm for VP1) and only minor transient fluctuations observed in VP1. RMSF analysis revealed localized flexibility, while radius of gyration, hydrogen bonding patterns, and solvent-accessible surface area collectively confirmed overall conformational stability throughout the simulation period. Complementary protein contact atlas and NCI analyses supported the persistence and robustness of protein–ligand interactions, showing comparable or improved stability relative to reference antivirals ribavirin and nitazoxanide. Additionally, all selected compounds exhibited favorable drug-likeness and acceptable ADME-Tox properties. Collectively, these findings suggest that the identified phytochemicals may serve as promising antiviral candidates against norovirus, although further validation through in vitro and in vivo studies is required.