Browse Articles
Discover research articles across all indexed journals
CBCT-to-CT synthesis using a hybrid U-Net diffusion model based on transformers and information bottleneck theory
A DEM super resolution reconstruction method based on normalizing flow
Publisher Correction: Eye movements of children with and without developmental dyslexia in an alphabetic script during alphabetic and logographic tasks
Experimental verification of the six sectors neural DTC approach of squirrel cage induction motors
Abstract The direct torque control (DTC) approach is one of the suitable solutions for controlling squirrel cage induction machines (SCIMs) due to its distinctive performance compared to other strategies and its simplicity. However, using this approach has several drawbacks and problems. This paper presents an experimental work using real equipment of an innovative method that combines six sectors of DTC technique and neural networks (NNs). The use of an NN algorithm allows for overcoming problems of the DTC approach, such as reducing torque ripples. Using the NN technique, the operation of the SCIM inverter is controlled, as the NN technique provides the pulses necessary to run the inverter, which allows for improving the quality of the current. Therefore, the presented approach is based on the usual method, using the same estimation equations. First, the validity of the designed approach was tested using MATLAB, comparing the results with the DTC approach. The results obtained showed a high ability of the six sectors’ NN-DTC approach to significantly enhance the quality of torque and current, which confirms the competence of using NNs. Secondly, real equipment was used to verify the simulation results and the extent of the efficiency and competence of the six sectors NN-DTC approach compared to the DTC technique in terms of improving the quality of current and torque. These experimental results obtained are of great value in the field of control, as they give a clear picture of the advantage of the six sectors of the NN-DTC approach in improving the features of the control system, which makes it more suitable for different applications in the future.
Machine learning for risk prediction of acute kidney injury in patients with diabetes mellitus combined with heart failure during hospitalization
Machine learning assisted Co3O4/NiO popsicle sticks-infused electrospun nanofibers for efficient oxygen evolution reaction
Abstract Wide range of noble metal free bimetallic and trimetallic based electrocatalysts have been synthesized to develop efficient oxygen evolution reaction (OER) systems to-date, however, to determine which metal part of bimetallic and trimetallic electrocatalysts plays a significant role in controlling OER efficacy remains very challenging. To address this issue, herein we have employed machine learning (ML) for the first time to determine OER efficacy controlling metal element, thus leading to the development of an optimized bimetallic electrocatalyst. Briefly, we have designed a novel, simple ML optimized sustainable OER electrocatalyst based on Co3O4/NiO popsicle sticks (CNPS) infused polyaniline/cellulose acetate (a biopolymer) (PNCA) electrospun nanofibers supported on nickel foam (NF). ML optimized CNPS infused PNCA (CNPS@PNCA) electrode offers maximum and homogenous exposition of active sites and shows high OER activity by exhibiting low onset potential (1.41 V vs. RHE), overpotential (237 mV at 10 mA cm−2) and Tafel slope of 62.1 mV dec−1. Additionally, it shows a better stability of more than 100 h and is consistent with the reported literature.
Automated liver and spleen segmentation for MR elastography maps using U-Nets
Abstract To compare pretrained and trained U-Nets for liver and spleen segmentation in multifrequency magnetic resonance elastography (MRE) magnitude images for automated quantification of shear wave speed (SWS). Seventy-two healthy participants (34 ± 11 years; BMI, 23 ± 2 kg/m2; 51 men) underwent multifrequency MRE at 1.5T or 3T. Volumes of interest (VOIs) of liver and spleen were generated from MRE magnitude images with mixed T2-T2* image contrast and then transferred to SWS maps. Pretrained and trained 2D and 3D U-Nets were compared with ground truth values obtained by manual segmentation using correlation analysis, intraclass correlation coefficients (ICCs), and Dice scores. For both VOI and SWS values, pairwise comparison revealed no statistically significant difference between ground truth and pretrained and trained U-Nets (all p ≥ 0.95). There was a strong positive correlation for SWS between ground truth and U-Nets with R = 0.99 for liver and R = 0.81–0.84 for spleen. ICC was 0.99 for liver and 0.90–0.92 for spleen, indicating excellent agreement for liver and good agreement for spleen for all U-Nets investigated. Dice scores showed excellent segmentation performance for all networks with the 2D U-Net achieving slightly higher values for the liver (0.95) and spleen (0.90), though the differences between the three tested U-Nets were minimal. The excellent performance we found for automated liver and spleen segmentation when applying 2D and 3D U-Nets to MRE magnitude images suggests that fully automated quantification of MRE parameters within anatomical regions is feasible by leveraging the previously unexploited anatomical information conveyed in MRE magnitude images.
Carbohydrate intake and activation of gastric acid secretion decrease gastric estrogen secretion
Genomic and metabolomic insights into the antimicrobial compounds and plant growth-promoting potential of Bacillus velezensis B115
Gender differences in L1 vertebral strength in adults 50+ using automated CT-based finite element analysis
Tailoring adversarial attacks on deep neural networks for targeted class manipulation using DeepFool algorithm
Abstract The susceptibility of deep neural networks (DNNs) to adversarial attacks undermines their reliability across numerous applications, underscoring the necessity for an in-depth exploration of these vulnerabilities and the formulation of robust defense strategies. The DeepFool algorithm by Moosavi-Dezfooli et al. (2016) represents a pivotal step in identifying minimal perturbations required to induce misclassification of input images. Nonetheless, its generic methodology falls short in scenarios necessitating targeted interventions. Additionally, previous research studies have predominantly concentrated on the success rate of attacks without adequately addressing the consequential distortion of images, the maintenance of image quality, or the confidence threshold required for misclassification. To bridge these gaps, we introduce the Enhanced Targeted DeepFool (ET DeepFool) algorithm, an evolution of DeepFool that not only facilitates the specification of desired misclassification targets but also incorporates a configurable minimum confidence score. Our empirical investigations demonstrate the superiority of this refined approach in maintaining the integrity of images and minimizing perturbations across a variety of DNN architectures. Unlike previous iterations, such as the Targeted DeepFool by Gajjar et al. (2022), our method grants unparalleled control over the perturbation process, enabling precise manipulation of model responses. Preliminary outcomes reveal that certain models, including AlexNet and the advanced Vision Transformer, display commendable robustness to such manipulations. This discovery of varying levels of model robustness, as unveiled through our confidence level adjustments, could have far-reaching implications for the field of image recognition. Our code is available at https://github.com/FazleLabib/et_deepfool
Poor sleep quality was associated with increased plasma aldosterone concentration in community dwellers, a cross-sectional study
Transcriptomic and metabolomic insights into flavor variations in wild and cultivated Agaricus bisporus
Green infrastructure site prioritization to improve urban flood resilience in Monterrey and Brussels using a decision support model
Abstract Green infrastructure (GI) has been increasingly associated with urban flood resilience as it provides benefits in protecting communities from flood dangers and improving socio-economic capabilities. In order to optimize the GI advantages, it is necessary to engage in strategic prioritizing of implementation areas, considering local conditions. Despite a growing interest in connecting GI and flood resilience, there is still a lack of strategic-oriented GI planning models aimed at enhancing urban flood resilience. This study has introduced the Flood Resilience-based Urban Green Infrastructure Site Priority (FRUGISP) model, which employs a GIS-based multi-criteria assessment to determine the urban regions that should be prioritized for the implementation of GI systems, based on their flood resilience levels. The model was used to map the priority areas in Monterrey, Mexico, and Brussels, Belgium. Despite their distinct features, both cities face flood challenges. The results showed regions of utmost importance based on the flood resilience index and land availability for GI implementation. The model has the potential to be applied to other urban areas grappling with flood issues, providing guidance to decision-makers in selecting high-priority locations for GI projects. This approach can effectively address the difficulties posed by urban floods, ensuring the resilience of urban areas.
Exploring the applicability of “One-Size-Fits-All” road transport decarbonization strategies: a participatory energy systems modeling comparison of urban and non-urban municipalities
Abstract Despite the key role that local authorities play in shaping energy policies and implementing action plans, their level of involvement has been insufficiently examined. This study aims to assess how different socio-geographical factors impact the adoption of fossil-free vehicle technologies and fuels for private cars, buses, and trucks. Using a participatory energy systems modeling approach, this study explores the cost-optimal decarbonization of road transport in four urban and non-urban Swedish municipalities. By collaborating with local authorities, socio-technical scenarios are modeled to reflect climate actions, resources and infrastructure availability, as well as travel patterns. The findings reveal a preference for lower upfront costs in urban areas with shorter trip distances, leading to a higher small-size battery electric vehicles (BEVs) share. Conversely, in non-urban areas with longer trip distances, fuel economy, fuel cost, as well as operation and maintenance costs outweigh upfront costs, increasing average-size BEVs share. Buses and trucks also experience a growing BEVs and fuel cell vehicles (FCEVs) share, driven by their typically high annual mileage. Biofuels play an intermediate role until BEVs and FCEVs are reduced in cost. Tailoring decarbonization strategies to local contexts is essential for maximum effectiveness, balancing national and regional climate goals with urban and non-urban challenges.