Search Results (511)

Monocular depth estimation is a central problem in computer vision with applications in robotics, augmented reality, and autonomous driving, yet the self-attention mechanisms used by modern Transformer architectures remain opaque. In this work, we integrate SVD-Inspired Attention (SVDA) into the Dense Prediction Transformer (DPT), introducing a spectrally structured attention formulation for dense prediction that decouples directional alignment from spectral modulation through a learnable diagonal matrix embedded in normalized query–key interactions. Experiments on KITTI and NYU-v2 show that SVDA preserves competitive predictive performance while enabling intrinsic interpretability: on KITTI, AbsRel improves from 0.058 to 0.056 and ${\delta }_1$ from 0.976 to 0.979, while on NYU-v2, AbsRel improves from 0.133 to 0.124 and ${\delta }_1$ from 0.865 to 0.872. This is achieved with only 0.01% additional parameters, at the cost of a measurable runtime overhead associated with the added normalization and spectral modulation. More importantly, SVDA enables six spectral indicators that quantify entropy, rank, sparsity, alignment, selectivity, and robustness, revealing consistent cross-dataset and depth-wise patterns in how attention organizes during training. These properties make the model easier to inspect and better suited to applications where transparency and reliability are important, such as robotics and autonomous navigation. Full article

Even when an object’s color is accurately reproduced in a colorimetrically reproduced image (CRI), the perceived material appearance does not necessarily match that of the original object. This mismatch remains a challenge for faithfully reproducing real-world appearance in digital media. In this study, we investigated how lightness and chroma modulation affect the perception of glossiness, transparency, and roughness. These three attributes were quantitatively correlated with physical surface properties and image features through a direct comparison between objects and images. Observers selected the images that best matched the material appearance of the physical samples for each attribute. Image features derived from the gray-level co-occurrence matrix (GLCM) and surface roughness parameters were analyzed to compare the selected images with the CRI. In the lightness experiment, observers consistently selected images with higher lightness than the CRI, which was accompanied by increased complexity in the luminance distribution. In the chroma experiment, images with higher chroma were preferred; however, changes in GLCM features were negligible. Notably, stimuli with small local luminance differences at the CRI required larger shifts in image features to achieve perceptual matching. These findings indicate that modulating the luminance distribution is crucial for aligning the perceived appearance between physical objects and their digital representations. Full article

Recently, more sustainable and biodegradable packaging materials have begun to attract attention in food packaging due to major, rising concerns related to plastic usage. This study aims to develop and characterize biodegradable food packaging materials, namely poly(lactic acid) (PLA) and poly(ε-caprolactone) (PCL) enriched with pomegranate peel extract (PoPE). Firstly, the optimal extract selected was a 24 h maceration of PoPE with 60% ethanol, after production with different solvents and methods. PLA- and PCL-based films were produced via melt compounding with the addition of PoPE at different concentrations (1, 3, 5 and 10%, w/w). FTIR confirmed that the PoPE did not modify the chemical backbones of PLA or PCL, with only a more pronounced O–H band in PCL, suggesting mainly non-covalent/physical interactions. UV–Vis spectroscopy showed tunable warm coloration and strong UV shielding with reduced transparency; for PLA ~3–5 wt.%, PoPE enabled near-complete UV blocking, while PCL achieved very high UV protection even at low loadings. PoPE improved toughness in PLA (3–5 wt.%) and maintained ductility in PCL (1–10 wt.%). PoPE-added PLA and PCL films maintained thermal stability up to 10 wt.% according to TGA results. DSC/XRD indicated a matrix-dependent crystallization response. PLA remained largely amorphous, whereas PoPE promoted PCL crystallinity without changing polymer crystal polymorphs. SEM images revealed homogenous dispersion of PoPE in the films. Full article

A continuous-time Markov chain framework is developed for service life prediction of building assets, and three formulations are compared: a homogeneous generator, a time-varying generator, and a fractional model. The framework delivers survival, density of absorption time, hazard, and mean time to absorption. For the homogeneous case, state trajectories are computed using matrix exponentials. The time-varying case is solved both by local exponential propagation on a time grid and by direct integration of the Kolmogorov equation. The fractional case is implemented in two independent ways, via a truncated series expansion and via an in-house routine for the Mittag-Leffler function, which also allows the direct evaluation of survival and hazard from the standard fractional relations while avoiding singular behaviour at the origin. This study shows that non-homogeneous rates accelerate deterioration relative to the homogeneous benchmark, whereas fractional dynamics reproduce early-time acceleration followed by a slow decline of the hazard, which is consistent with heavy-tailed survival and longer effective service life. The two fractional solvers provide mutually consistent outputs, which supports the numerical robustness of the approach. The framework is readily applicable to sparse inspection data and short observation windows and provides a transparent basis for comparing modelling assumptions that affect life cycle forecasts used in asset management and maintenance planning. Full article

We develop a theoretical description of the electrophoretic migration of a weakly charged oil drop dispersed in a dilute polymer gel carrying fixed charges and saturated with an aqueous electrolyte solution. In contrast to neutral gels, a charged polymer network generates electroosmotic flow under an applied electric field, which couples with the electrohydrodynamic motion of the drop. The observed electrophoretic velocity therefore reflects the combined effects of drop-induced flow and gel-driven electroosmosis. On the basis of the Baygents–Saville theory, the drop surface charge is assumed to originate from specific ion adsorption at the oil–water interface, while no mobile ions are present inside the drop. Working within the Brinkman–Debye–Bueche porous-medium model for the gel and employing a linearized treatment valid for low zeta potential, we obtain a simple analytical expression for the electrophoretic mobility. The formulation consistently incorporates Marangoni stresses arising from spatial variations in interfacial tension, and hydrodynamic slip at the oil–water interface, which can be significant for hydrophobic drops in aqueous media. The resulting mobility expression explicitly separates the contribution associated with the intrinsic electrohydrodynamic response of the drop from that due to electroosmosis of the charged gel matrix. This analytical form enables experimental mobility data to be used not only to estimate the zeta potential of the drop but also to evaluate the electroosmotic mobility of the polymer gel medium. The present theory thus provides a physically transparent and experimentally useful framework for characterizing electrokinetic transport in charged soft porous media. Full article

Background: Geographic and demographic disparities strongly influence access to community-based healthcare, especially in rural and mountainous areas. In Italy, Ministerial Decree 77/2022 promotes a territorial reorganization based on networked care models, but practical tools for translating policy standards into operational catchment areas remain limited. Methods: We developed a transparent, data-driven allocation framework combining travel-time accessibility and population-based capacity constraints. A case study was conducted in the Province of L’Aquila, within Local Health Authority ASL 1 Avezzano–Sulmona–L’Aquila, a low-density mountainous area including 65 municipalities. Using official ISTAT data, including the 2021 national origin–destination road travel-time matrix, municipalities were allocated to 3 hub nodes and 8 spoke nodes. Population caps of 50,000 residents per hub and 40,000 per spoke were applied. Scenario analyses were performed under 20, 30, and 40 min travel-time thresholds. Results: Under the 30 min scenario, all municipalities were allocated, but the L’Aquila hub exceeded the capacity cap. A cap-compliant 30 min allocation eliminated this violation at the cost of longer upper-tail travel times. Under the 20 min scenario, only 54 municipalities were allocated, leaving 11 mountainous municipalities outside the threshold. Under the 40 min scenario, all municipalities were allocated without capacity violations. Conclusions: The proposed framework provides a reproducible approach for territorial healthcare planning and makes explicit the trade-off between accessibility and capacity compliance in hub-and-spoke network design, particularly in geographically complex mountain settings. Full article

Protein-based films are promising biodegradable materials, but their performance is often limited by structural instability during storage. In this study, blend films were developed from myofibrillar proteins of giant squid (Dosidicus gigas) and whey protein concentrate (WPC) to improve functional properties and evaluate stability during three months of storage. The effects of plasticizer type (glycerol or sorbitol) and WPC concentration (5–15%) on film structure and performance were analyzed using Fourier Transform Infrared Spectroscopy (FT-IR), Thermogravimetric Analysis (TGA), optical measurements, solubility, and water vapor transmission rate (WVTR). FT-IR revealed a transition from α-helix to β-sheet structures, indicating stronger protein–protein interactions, particularly in sorbitol-plasticized films. This structural organization improved barrier properties, reducing WVTR from 44.2 g·m⁻²·d⁻¹ in squid protein films to 18.9 g·m⁻²·d⁻¹ in films containing WPC. Light transmittance analysis showed that all films acted as effective UV barriers, with transmission starting near 350 nm. At this wavelength, transmittance ranged from 5–17% in sorbitol-plasticized films to 33–46% in glycerol-plasticized films. Increasing WPC concentration also reduced film solubility, indicating the formation of a more compact protein matrix. During three months of storage, FT-IR spectra revealed changes in the Amide A and Amide III bands associated with plasticizer migration and increased protein–protein interactions. Transparency increased during storage, indicating progressive structural reorganization, while the UV barrier properties remained stable. These results demonstrate that blending squid and whey proteins, particularly with sorbitol as plasticizer, produces biodegradable films with improved barrier properties and good structural stability during storage, highlighting their potential for sustainable food packaging applications. Full article

Artificial intelligence (AI) has significantly advanced predictive capabilities in digital health systems; however, the structural integration of these predictions into formal decision-making processes remains insufficiently addressed. This study introduces the Decision Integration Deficit Index (DIDI), a structural diagnostic metric designed to assess the alignment between inference- and decision-oriented components in AI-driven health system architectures. A domain × domain integration matrix represents structurally possible and empirically observed relationships between system components, enabling the formal assessment of integration patterns. The framework suggests that apparent balance at an aggregated level may conceal substantial structural asymmetries, particularly in the limited integration of modelling outputs into formal evaluation and decision-support mechanisms. The results suggest that the analyzed corpus reflects structurally incomplete architectures, characterized by an imbalance in decision integration across domains. In contrast to performance-based evaluation metrics, the DIDI provides a system-level diagnostic perspective that identifies missing or weakly specified integration pathways within decision-process architectures. This study contributes to digital health and decision-support research by introducing a reproducible structural assessment framework that enables evaluation of decision-process completeness and supports the development of more coherent, transparent, and accountable AI-driven decision-support systems. Full article

Building Information Modelling (BIM) offers a pivotal opportunity to automate Life Cycle Assessment (LCA) within the Architecture, Engineering, and Construction (AEC) industry. However, seamless integration is persistently hindered by a semantic gap, a critical misalignment between the object-oriented, geometric definitions of BIM and the process-based material data required by Life Cycle Inventory (LCI) databases. This paper presents a comprehensive review of data interoperability challenges and evaluates advanced methodologies designed to bridge this divide, moving beyond simple tool comparison to analyse structural integration barriers. Through a systematic review of 124 primary studies published between 2010 and 2025, this research inductively derives the BIM-LCA Interoperability Triad. This framework analyses causal dependencies across three dimensions, including Semantic and Ontological Structures, Workflow and Temporal Integration, and System Architecture and Interoperability. Furthermore, by establishing a comparative challenge–solution matrix, the analysis reveals a maturity paradox in current methodologies. While semi-automated commercial plugins dominate practice due to accessibility, they frequently function as opaque black boxes with limited transparency. Conversely, advanced approaches utilising Semantic Web technologies and Machine Learning demonstrate superior capability in resolving terminological mismatches but currently face significant barriers regarding infrastructure and expertise. This study contributes a novel theoretical model for understanding integration failures. It concludes that future research must pivot from static schema mapping towards AI-driven semantic healing, dynamic Digital Twins, and explicit system boundary harmonisation to achieve truly automated, context-aware environmental assessments and support whole-life circularity. Full article

Corneal fibrosis, clinically referred to as corneal scarring, disrupts the normal architecture and transparency of the cornea and remains a major cause of visual impairment worldwide. Although corneal transplantation can restore vision, its effectiveness is constrained by limited accessibility, donor tissue shortages, and the risk of allograft rejection. Treatments with human corneal stromal stem cells (hCSSCs) have demonstrated scarless healing in preclinical models of acute corneal injury. Here, we report that hCSSCs also modulated pre-existing corneal opacities. We established a reproducible in vivo model of chronic corneal opacity. Given that scar severity varies among corneas even after identical injuries, we developed a non-invasive, image-based method to quantify opacity volume longitudinally in individual corneas. Using this approach, we evaluated the scar-reducing potential of three hCSSC batches previously shown to inhibit acute scarring. Following cell treatment, the pre-existing opacity volumes gradually decreased. In vitro, hCSSCs exposed to pro-inflammatory stimulus exhibited increased metalloproteinase (MMP) activity relative to tissue inhibitor of metalloproteinase (TIMP), as indicated by an elevated MMP2/TIMP2 ratio. This shift may promote matrix remodeling and scar resolution. Overall, our findings provide proof-of-concept for hCSSC-based therapy as a strategy to reduce established corneal scarring and restore corneal transparency. Full article

This research focuses on developing bioplastic films using agrifood industrial waste, which included starch from avocado seed, cellulose from cornstalk, carrot and beet peel, and pulp from a food company in México. The films were produced with a matrix of gelatin and glycerol, and different formulations of starch and cellulose. The films were characterized and tested as wrappers of Manchego cheese. The films containing starch are transparent; films with cellulose showed opacity and paper-like structure. Films containing starch–cornstalk cellulose showed the highest hydrophobic properties. In turn, films with carrot cellulose had the highest plastic properties with high elongation capacity and the lowest Young modules; films with starch and other celluloses showed the opposite data. The highest thermal capacity was observed in films containing cellulose from cornstalks and beet waste. In turn, the highest temperatures of transition, crystallization, and melting were registered in films containing starch. Films with starch and cellulose served well as wrappers of Manchego cheese, conserving 92% of the weight of cheese after 21 days of storage at 4 °C. All films were biodegradable in compost after 10 days, and they were degradable by physicochemical factors after 40 days. Full article

This study investigates the effects of organic salts, including sodium citrate (SC), calcium citrate (CC), and calcium lactate (CL), on the structure–property–function relationships of thermoplastic starch/poly(butylene adipate-co-terephthalate) (TPS/PBAT) films for active packaging applications. TPS incorporated with organic salts was prepared via twin-screw extrusion, blended with PBAT, and further processed into blown films. The films were systematically characterized using ¹H NMR, FTIR, and SEM, together with optical, mechanical, water vapor permeability, and antimicrobial evaluations against Staphylococcus aureus. The results revealed that SC primarily modulated hydrogen-bonding interactions within the starch matrix, resulting in improved structural homogeneity, balanced mechanical properties, and the highest antimicrobial activity among all formulations. In contrast, CL and CC promoted ionic crosslinking through Ca²⁺–starch interactions, leading to increased stiffness and Young’s modulus but reduced polymer chain mobility and limited release of active species, particularly in CC-containing systems. These differences in molecular interactions were consistent with variations in film microstructure, where SC-containing films exhibited more uniform morphologies, while calcium-based systems showed denser but less permeable structures. Furthermore, films containing SC and CL at appropriate concentrations achieved a favorable balance between transparency, water vapor barrier properties, and antimicrobial performance. Overall, this study provides new mechanistic insights into how monovalent and divalent organic salts govern intermolecular interactions, microstructure, and functional performance in TPS/PBAT systems. The findings highlight the critical role of additive type and concentration in designing biodegradable active packaging materials with tunable mechanical, barrier, and antimicrobial properties. Full article

Continuous unidirectional (UD) thermoplastic composite tapes are increasingly used in aerospace, automotive, and energy applications because of their high specific strength, low weight, recyclability, and compatibility with automated manufacturing. Since final component performance strongly depends on tape quality, reliable defect characterization and detection are essential. In this study, manufacturing-induced defects in polypropylene-based UD tapes reinforced with carbon and glass fibers were investigated using real images acquired directly from laboratory-scale production without synthetic data. Defects related to interfacial integrity, matrix distribution, fiber architecture, and surface irregularities were systematically analyzed, and a practical four-class defect taxonomy was established. To enable automated inspection under limited-data conditions, lightweight YOLOv8, YOLOv11, and the new YOLO26 models were comparatively evaluated using a UD tape-specific augmentation strategy combining physically constrained Albumentations and on-the-fly augmentation. Among the tested models, YOLO26-s achieved the best overall performance, reaching a mean mAP@0.5 of 0.87 ± 0.03, outperforming YOLOv11 (0.83) and YOLOv8 (0.78), with 0.90 precision and 0.85 recall. Interfacial (0.92 mAP) and matrix-related (0.90 mAP) defects were detected most reliably, whereas fiber-related (0.89 mAP) and surface defects (0.79 mAP) remained more challenging, particularly in glass-fiber-reinforced tapes due to transparency-masking effects. The results demonstrate the potential of compact deep learning models for computationally efficient and manufacturing-relevant in-line quality monitoring of UD tape production. Full article

This study developed an active packaging material by incorporating tea tree oil (TTO)-loaded lotus stalk biochar (BC@TTO) into a chitosan (CS) matrix. Biochar was prepared from lotus stalks via pyrolysis at 600 °C and characterized, revealing a mesoporous structure with a specific surface area of 35.9 m²/g. Adsorption studies demonstrated that BC exhibited high affinity for TTO, following pseudo-first-order kinetics and the Langmuir isotherm model, with a maximum adsorption capacity of 295.6 mg/g. Chitosan-based composite membranes with varying BC@TTO contents (1–7 wt%) were fabricated by solution casting. The incorporation of BC@TTO significantly enhanced the tensile strength, elongation at break, barrier properties (water vapor and oxygen), and antioxidant/antibacterial activities of the membranes, with optimal performance observed at 3 wt% loading. However, higher loadings led to filler aggregation, reduced transparency, and compromised mechanical properties. In vitro release studies indicated that TTO release followed the Avrami model, suggesting a diffusion-controlled mechanism. Preservation tests on blueberries showed that the CS-3BC@TTO membrane effectively reduced weight loss and maintained fruit quality during storage. This work presents a promising strategy for designing bioactive packaging materials with sustained release functionality for food preservation applications. Full article

This article details a matrix-based mathematical method to calculate power flows and transmission losses in an electric grid specifically attributable to low-carbon and renewable energy sources (LCRES) (wind, solar, nuclear). The goal is to improve the transparency and reliability of Guarantees of Origin (GO) certificates. Current GO schemes rely on contractual accounting and neglect physical power losses, undermining consumers’ confidence that they receive “clean” energy. The method uses steady-state power flow analysis to derive a power-loss distribution coefficient matrix. This matrix accurately allocates grid losses back to the LCRES generating nodes, complying strictly with electrical engineering principles. It accommodates both time-varying renewable output and stable nuclear generation. The results offer highly accurate loss-attribution data, supporting more verifiable GOs, ensuring fair compensation for losses, and enhancing energy balance accuracy in hybrid power systems. Full article