Search Results (3,781)

Background/Objectives: Accurate volumetry of the remnant pancreas after pancreatectomy is crucial for assessing postoperative endocrine and exocrine function but remains challenging due to anatomical variability and complex postoperative morphology. This study aimed to develop and validate a deep learning (DL)-based model for automatic segmentation and volumetry of the remnant pancreas using abdominal CT images. Methods: A total of 1067 CT scans from 341 patients who underwent pancreaticoduodenectomy and 512 scans from 184 patients who underwent distal pancreatectomy were analyzed. Ground truth masks were manually delineated and verified through multi-expert consensus. Six 3D segmentation models were trained and compared, including four convolution-based U-Net variants (basic, dense, residual, and residual dense) and two transformer-based models (Trans U-Net and Swin U-Net). Model performance was evaluated using five-fold cross-validation with sensitivity, specificity, precision, accuracy, and Dice similarity coefficient. Results: The Residual Dense U-Net achieved the best performance among convolutional models, with dice similarity coefficient (DSC) values of 0.7655 ± 0.0052 for pancreaticoduodenectomy and 0.8086 ± 0.0091 for distal pancreatectomy. Transformer-based models showed slightly higher DSCs (Swin U-Net: 0.7787 ± 0.0062 and 0.8132 ± 0.0101), with statistically significant but numerically small improvements (p < 0.01). Conclusions: The proposed DL-based approach enables accurate and reproducible postoperative pancreas segmentation and volumetry. Automated volumetric assessment may support objective evaluation of remnant pancreatic function and provide a foundation for predictive modeling in long-term clinical management after pancreatectomy. Full article

Wall-thickness deformation is a critical indicator of fatigue risk in flexible risers exposed to vortex-induced vibration (VIV), especially under combined internal and external flow conditions. This study examines the spanwise evolution and distribution of wall-thickness deformation in a riser traversing air and water. The effects of external flow velocity, internal flow velocity, and internal fluid density on in-line (IL) and cross-flow (CF) wall deformation are systematically analyzed at characteristic positions. The results show that wall deformation exhibits strong spatial variability and media property dependence: IL deformation in the air-exposed segment is amplified under lock-in conditions due to lower damping, while the submerged segment experiences consistently larger deformation driven by added-mass effects. Internal flow influences wall-thickness response in a non-monotonic manner, and increased internal fluid density suppresses deformation while shifting the dominant frequency. These findings demonstrate that wall-thickness deformation is a sensitive and integrative response to fluid–structure interaction, offering a direct basis for identifying high-risk zones and improving fatigue-resistant design in deep-sea riser systems. Full article

The iron and steel industry is one of the largest industrial sources of greenhouse gas emissions. This paper examines the potential of green hydrogen as a reducing agent for decarbonizing primary steel production, focusing on the Taranto integrated steelworks in southern Italy. Producing about 3.5 Mt of crude steel annually, the plant is also among the country’s biggest emitters, with CO₂ emissions of roughly 8 Mt per year at typical blast furnace intensity (2.2 tCO₂/t steel). The analysis quantifies the hydrogen demand required to replace fossil fuels in iron ore reduction and evaluates the techno-economic feasibility of meeting it with green hydrogen. Using DWSIM (open-source chemical process simulation software, v9.0.2) for water electrolysis powered by renewables, the study estimates both the CO₂ emission reductions and cost impacts of hydrogen-based steelmaking. Results show that integrating green hydrogen at Taranto could achieve deep decarbonization by cutting emissions by over 90%, with a base-case levelized hydrogen cost (LCOH) of 3.6 EUR/kg and green steel production cost 653 EUR/t. With optimistic assumptions (renewable electricity at 40 EUR/MWh and electrolyzer CAPEX halved to 500 EUR/kW), hydrogen cost could be reduced to 2.3 EUR/kg, making green steel cost-competitive with conventional steel and implying a breakeven carbon price of under 60 EUR/t. Sensitivity analyses highlight that falling renewable electricity prices, supportive carbon policies, and successful demonstration projects are key enablers for economic viability. The findings underscore that renewable hydrogen can be a viable decarbonization pathway for steel when coupled with continued technological improvements and policy support. Full article

Background/Objectives: Sarcopenia and muscle composition have emerged as significant indicators in the fields of musculoskeletal and metabolic research. The objective of this study was to develop and validate a fully automated, deep learning-based method for segmenting thigh muscles into three functional groups (extensor, flexor, and adductor) using non-contrast computed tomography (CT) images and to quantitatively evaluate the thigh muscles. Methods: In order to ascertain the most efficacious architecture for automated thigh muscle segmentation, three deep learning models (Dense U-Net, MANet, and SegFormer) were implemented and subsequently compared. Each model was trained using 136 manually labeled non-contrast thigh CT scans and externally validated with 40 scans from another institution. The performance of the segmentation was evaluated using the Dice similarity coefficient (DSC), sensitivity, specificity, and accuracy. Quantitative indices, including total muscle volume, lean muscle volume, and intra-/intermuscular fat volumes, were automatically calculated and compared with manual measurements. Results: All three models exhibited high segmentation accuracy, with the mean DSC exceeding 96%. The MANet model demonstrated optimal performance in internal validation, while the SegFormer model exhibited superior volumetric agreement in external validation, as indicated by an intraclass correlation coefficient (ICC) of at least 0.995 and a p-value less than 0.01. Conclusions: A CT-based deep learning framework enables accurate and reproducible segmentation of functional thigh muscle groups. A comparative evaluation of convolutional attention- and transformer-based architectures supports the feasibility of CT-based quantitative muscle assessment for sarcopenia and musculoskeletal research. Full article

Accurate stock price forecasting is crucial for supporting informed investment decisions, effective risk management, and the identification of profitable market opportunities. Financial time series present considerable challenges for prediction due to their complex, nonlinear dynamics and sensitivity to a wide range of economic factors. Although various statistical methods have been developed to model the multidimensional relationships inherent in such datasets, advancements in big data technologies have greatly facilitated the recording, analysis, and interpretation of large-scale financial data, thereby accelerating the adoption of deep learning (DL) algorithms in this domain. In the present study, RNN-, LSTM-, and GRU-based models were developed to forecast the closing prices of two airline stocks, with hyperparameter optimization conducted via the Bayesian optimization algorithm. The dataset consisted of daily closing prices of THYAO and PGSUS stocks obtained from Yahoo Finance. Comparative analysis demonstrated that the GRU model yielded the highest accuracy for THYAO stock price prediction, achieving a MAPE of 3.05% and an RMSE of 3.195, whereas for PGSUS, the model achieved a MAPE of 3.97% and an RMSE of 3.232. Beyond its empirical contribution, this study also emphasizes the conceptual relevance of symmetry in financial forecasting. The proposed deep learning framework captures the balanced relationships and nonlinear interactions inherent in stock market behavior, reflecting both symmetry and asymmetry in market responses to economic factors. Full article

We propose a novel hybrid deep learning framework that synergistically integrates Convolutional Neural Networks (CNNs), Spiking Neural Networks (SNNs), and Generative Adversarial Networks (GANs) for robust and accurate classification of high-resolution frontal and sagittal human gait video sequences—capturing both lower-limb kinematics and upper-body posture—from subjects with Knee Osteoarthritis (KOA), Parkinson’s Disease (PD), and healthy Normal (NM) controls, classified into three disease-type categories. Our approach first employs a tailored CNN backbone to extract rich spatial features from fixed-length clips (e.g., 16 frames resized to 128 × 128 px), which are then temporally encoded and processed by an SNN layer to capture dynamic gait patterns. To address class imbalance and enhance generalization, a conditional GAN augments rare severity classes with realistic synthetic gait sequences. Evaluated on the controlled, marker-based KOA-PD-NM laboratory public dataset, our model achieves an overall accuracy of 99.47%, a sensitivity of 98.4%, a specificity of 99.0%, and an F1-score of 98.6%, outperforming baseline CNN, SNN, and CNN–SNN configurations by over 2.5% in accuracy and 3.1% in F1-score. Ablation studies confirm that GAN-based augmentation yields a 1.9% accuracy gain, while the SNN layer provides critical temporal robustness. Our findings demonstrate that this CNN–SNN–GAN paradigm offers a powerful, computationally efficient solution for high-precision, gait-based disease classification, achieving a 48.4% reduction in FLOPs (1.82 GFLOPs to 0.94 GFLOPs) and 9.2% lower average power consumption (68.4 W to 62.1 W) on Kaggle P100 GPU compared to CNN-only baselines. The hybrid model demonstrates significant potential for energy savings on neuromorphic hardware, with an estimated 13.2% reduction in energy per inference based on FLOP-based analysis, positioning it favorably for deployment in resource-constrained clinical environments and edge computing scenarios. Full article

The oil layers in the Dongying Formation offshore oilfield are severely contaminated. The near-wellbore reservoir pore throats are blocked, which seriously affects the development effect. It has become urgent to implement acidizing stimulation measures. However, the target reservoir is deeply buried, has high reservoir pressure, and is highly sensitive. These factors result in high pressure during acidizing operations, a long single-trip time for raising and lowering the tubing string, and high costs. Moreover, acid that is not promptly returned to the surface after acidizing can cause secondary pollution to the reservoir. This work proposes an integrated tubing string to perform reverse displacement and reverse squeeze. With this, acid can be injected into the formation through the annulus between the casing and tubing. The residual acid and its post-acidizing derivative residues are rapidly lifted to the surface by the reciprocating suction action of the return pump. Based on this, the structure and specifications of the acidizing-flowback tubing string are designed through the flow rate analysis method. The tubing string is mainly affected by mechanical effects, including buoyancy, piston effect, flow viscosity effect, helical bending effect, temperature difference effect, and expansion effect. The maximum deformations are 1.4 m, 1.9 m, 0.18 m, 2.7 m, 1.8 m, and 2.5 m, respectively. The total deformation is less than 3 m. Simulation results from three groups of oil wells at different depths indicate that the axial force of the tubing string ranges from 400 to 600 kN. The stress ranges from 260 to 350 MPa, deformation is 1.1–2.4 mm, and the safety factor exceeds 3.0. This can effectively ensure the safety of on-site operations. Based on the actual field conditions, the acidizing-flowback tubing string is evaluated. This verifies the effectiveness of the acidizing-flowback tubing string. This research provides an economical and efficient operation process for acidizing operations in the Dongying Formation offshore oilfield. It achieves the goal of removing reservoir contamination and provides guidance for the unblocking and stimulation of low-permeability and sensitive reservoirs in the middle and deep layers of offshore oilfields. Full article

The Metaverse—an evolving convergence of virtual and physical realities—has emerged as a transformative platform, particularly within the fashion and retail industries. Its immersive nature aligns closely with the principles of flow theory, which describes the optimal psychological state of deep engagement and enjoyment. This study investigates the dynamics of fashion retail shopping in the Metaverse through a novel multi-criteria decision-making (MCDM) framework. Specifically, it integrates the CRITIC and ARAS methods within a spherical fuzzy environment to address decision-making under uncertainty. Flow theory is employed as the theoretical lens, with its dimensions serving as evaluation criteria. By incorporating spherical fuzzy sets, the model accommodates expert uncertainty more effectively. The findings provide empirical insights into the relative importance of flow constructs in shaping immersive consumer experiences in Metaverse-based retail environments. This study offers both theoretical contributions to the literature on digital consumer behavior and practical implications for fashion brands navigating immersive virtual ecosystems. Sensitivity analyses and comparative validation further demonstrate the robustness of the proposed framework. Full article

Chinese shale reservoirs are typically deep, clay-rich, and highly water-sensitive, which severely limits the effectiveness of conventional hydraulic fracturing. To address this challenge, we propose a microwave-assisted waterless fracturing method and investigate its feasibility through laboratory experiments on core samples from the Gulong shale and tight sandstone formations in the Daqing Oilfield. A coupled model integrating microwave power dissipation, pore water heating, and thermal stress evolution is developed to interpret the underlying mechanisms. Experimental results show that, under microwave irradiation (200 W, 40 s) and initial pore water content of 2.1–6%, fracturing is successfully induced without external fluid injection. The tensile failure of the rock coincides with the peak internal pore pressure generated by rapid vaporization and thermal expansion of pore water, as confirmed by a modified tensile strength measurement method. Fracture patterns observed in SEM and post-treatment imaging align with model predictions, demonstrating that microwave energy absorption by pore water is the primary driver of rock failure. The technique eliminates water-related formation damage and is inherently suitable for deep, water-sensitive reservoirs. This study provides experimental evidence and mechanistic insight supporting microwave-based waterless fracturing as a viable approach for challenging shale formations. Full article

Background: A significant concern today is the dependence on low-quality water sources, such as saline water, in hydroponic systems, especially due to the scarcity of freshwater. Halophytes and salt-tolerant species have emerged as viable candidates for cultivation in saline hydroponics. However, their agronomic performance and physiological responses within hydroponic systems require further investigation. Objectives: This research aims to explore the potential of edible halophytes grown in saline nutrient solutions within hydroponic systems within salt-tolerant ranges, focusing on their metabolic profiles and mineral accumulation. Methods: Plantago coronopus (L.), Portulaca oleracea (L.), and Salsola komarovii (Iljin) were grown in walk-in controlled environment chambers in deep water culture hydroponic systems, at 0, 50, 100, 150, and 200 mM·L⁻¹ NaCl salinity; 16h, 250 µmol m⁻² s⁻¹, and wide LED spectrum lighting was maintained. Results: A significant decrease in organic acids, and fresh and dry weight under high saltinity was observed in Plantago coronopus and Portulaca oleracea, but not in Salsola komarovii. An increase in hexoses, particularly glucose, violaxanthin and β-carotene, P⁺ and Zn²⁺, along with a decrease in lutein, K⁺ and Ca²⁺ levels across salinity levels from 0 to 200 mM NaCl was observed in all treated halophytes. Increased salinity did not significantly affect total protein accumulation. Conclusions: These findings reveal that different shifts in osmolytes, mineral elements, and biomass accumulation in tested halophytes indicate species-dependent osmotic adjustment to increased salinity and may be attributed to the morphological differences among halophytic grasses, dicot halophytes, and those with succulent leaves or stems. The PCA score scatterplot results excluded the response of Plantago coronopus from other tested halophytes; also, it demonstrated that Portulaca oleracea was more sensitive to the hydroponic solution salinity compared to Salsola komarovii and Plantago coronopus. Full article

Background: While gravity-assisted ankle stress AP (GAASA) images have proven effective in evaluating deep deltoid ligament injuries, their efficacy in assessing syndesmosis injuries remains unclear. We aimed to investigate the diagnostic performance of GAASA images in detecting syndesmosis injuries. Methods: This study reviewed records of patients aged 16+ with unilateral ankle fractures in a single-center ER from 2022 to 2023. Three orthopedic surgeons evaluated standard AP and lateral X-rays, ankle mortise, and GAASA and bilateral ankle CT images in blinded sessions for syndesmosis injuries. Evaluations were repeated to assess the inter- and intra-rater reliability. Results: A total of 121 patients with suspected syndesmosis injuries were included in this study. The average age of the patients was 49.9 ± 16.6 years. Syndesmosis injuries were present in 32.2% of cases. The inter-observer reliability was the highest for GAASA images (κ = 0.701) and mortise radiographs (κ = 0.735), and lowest for CT images (κ = 0.426). GAASA images had a sensitivity of 82% and specificity of 68%. Mortise images had 55% sensitivity and 81% specificity. GAASA images showed better discriminatory power for syndesmosis injuries compared to mortise and CT images. Conclusions: GAASA images demonstrated superior sensitivity and better negative predictive values in detecting syndesmosis injuries compared to mortise radiographs and CT images. While GAASA may serve as a useful adjunct for evaluating syndesmosis injuries, its interpretation requires careful clinical correlation, and it should not be considered a replacement for standard imaging in all cases. GAASA may be of particular value in emergency or resource-limited settings where CT is not readily available, offering a practical option for ruling out injury in many patients. Full article

Magnetic Particle Imaging (MPI) is a cutting-edge noninvasive imaging technique that offers high sensitivity, quantitative accuracy, and operates without the need for ionizing radiation compared to other imaging techniques. Utilizing superparamagnetic iron oxide nanoparticles (SPIONs) as tracers, MPI enables direct and precise visualization of target sites with no limitation on imaging depth. Unlike magnetic resonance imaging (MRI), which relies on uniform magnetic fields to produce anatomical images, MPI enables direct, background-free visualization and quantification of SPIONS within living organisms. This article provides an in-depth overview of MPI’s applications in tracking tumor development and supporting cancer therapy. The distinct physical principles that underpin MPI, including its ability to produce high-contrast images devoid of background tissue interference, facilitating accurate tumor identification and real-time monitoring of treatment outcomes, are outlined. The review outlines MPI’s advantages over conventional imaging techniques in terms of sensitivity and resolution, and examines its capabilities in visualizing tumor vasculature, tracking cellular movement, evaluating inflammation, and conducting magnetic hyperthermia treatments. Recent progress in tracer optimization and magnetic navigation has expanded MPI’s potential for targeted drug delivery, along with deep machine learning procedures for MPI applications. Additionally, considerations around safety and the feasibility of clinical implementation are also discussed in the present review. Overall, MPI is positioned as a promising tool in advancing cancer diagnostics, personalized therapy assessment, and noninvasive treatment strategies. Full article

The purpose of the study is to investigate human physiological responses to construction noise exposure using deep learning, applying electroencephalography (EEG) and electro-dermal activity (EDA) sensors. Construction noise is a pervasive occupational stressor that affects physiological states and impairs cognitive performance. EEG sensors capture neural activity related to perception and attention, and EDA reflects autonomic arousal and stress. In this study, twenty-five participants were exposed to impulsive noise from pile drivers and tonal noise from earth augers at three intensity levels (40, 60, and 80 dB), while EEG and EDA signals were recorded simultaneously. Convolutional neural networks (CNN) were utilized for EEG and long short-term memory networks (LSTM) for EDA. The results depict that EEG-based models consistently outperformed EDA-based models, establishing EEG as the dominant modality. In addition, decision-level fusion enhanced robustness across evaluation metrics by employing complementary information from EDA sensors. Ablation analyses presented that model performance was sensitive to design choices, with medium EEG windows (6 s), medium EDA windows (5–10 s), smaller batch sizes, and moderate weight decay yielding the most stable results. Further, retraining with ablation-informed hyperparameters confirmed that this configuration improved overall accuracy and maintained stable generalization across folds. The outcome of this study demonstrates the potential of deep learning to capture multimodal physiological responses when subjected to construction noise and emphasizes the critical role of modality-specific design and systematic hyperparameter optimization in achieving reliable annoyance detection. Full article

Background/Objectives: Long-read metagenomic sequencing can detect bacteria and antimicrobial resistance genes (ARGs) from bovine respiratory samples, providing an alternative to culture and antimicrobial susceptibility testing (C/S). This study applied Bayesian latent class models (BLCMs) to estimate the sensitivity (Se) and specificity (Sp) of long-read metagenomic sequencing compared to C/S for detecting Mannheimia haemolytica, Pasteurella multocida, and Histophilus somni, as well as associated macrolide and tetracycline resistance potential. Methods: Deep nasopharyngeal swabs were collected from fall-placed feedlot calves at arrival, 13, and 36 days on feed across two years and two metaphylaxis protocols. Samples underwent C/S and long-read metagenomic sequencing. BLCMs were used to estimate Se and Sp for the detection of bacteria and potential for antimicrobial resistance (AMR). Results: Se and Sp for detecting respiratory bacteria by metagenomics were not significantly different than culture, with four exceptions. For the 2020 samples, Se for M. haemolytica was lower than culture, and Sp for H. somni was lower, while in both 2020 and 2021 samples, Se for P. multocida was higher for metagenomics than culture. The estimated Se and Sp of metagenomics for the detection of msrE-mphE, EstT, and tet(H) within bacterial reads were either not significantly different or were lower than AST, with Sp > 95% with one exception. Conclusions: This study provided BLCM-based estimates of clinical Se and Sp of metagenomics and C/S without assuming a gold standard in a large pen research setting. These findings demonstrate the potential of long-read metagenomics to support bovine respiratory disease diagnostics, AMR surveillance, and antimicrobial stewardship in feedlot cattle. Full article

Permeability evolution is one of the key parameters influencing the efficient exploitation of deep unconventional energy resources, as it reflects the dynamic development of pore-fracture structures under complex engineering effects. Using fractal geometry to describe the pore-fracture system, rock permeability enhancement can be quantitatively evaluated. In this study, fractured coal specimens were analyzed under simulated mining-induced stress relief and CH₄ release conditions based on fractal geometry theory. The permeability-enhancement rate was derived and verified through CT (Computed Tomography) characterization of the pore-fracture network. The fractal dimension of the fracture aperture distribution and the tortuosity of fracture paths were determined to establish a fractal permeability-enhancement model, and its sensitivity was analyzed. The results indicate that permeability evolution undergoes four distinct stages: a stable stage, a slow-growth stage, a rapid-growth stage, and a stable or declining stage. The mining-induced stress relief and gas desorption effects significantly accelerate permeability enhancement, providing new insights into the mechanisms governing gas flow and pressure relief in deep coal seams. The proposed model, highly sensitive to the fracture aperture ratio (λ_min/λ_max), reveals that a smaller aperture span leads to greater permeability enhancement during the damage and fracture stage. These findings offer practical guidance for predicting permeability evolution, optimizing gas drainage design, and enhancing the safety and efficiency of coal mining and methane extraction operations. Full article