A lightweight deep learning approach for COVID-19 detection using X-ray images with edge federation

Article Subjects > Biomedicine
Subjects > Engineering
Europe University of Atlantic > Research > Scientific Production
Fundación Universitaria Internacional de Colombia > Research > Scientific Production
Ibero-american International University > Research > Articles and books
Ibero-american International University > Research > Scientific Production
Universidad Internacional do Cuanza > Research > Scientific Production
Abierto Inglés Objective This study aims to develop a lightweight convolutional neural network-based edge federated learning architecture for COVID-19 detection using X-ray images, aiming to minimize computational cost, latency, and bandwidth requirements while preserving patient privacy. Method The proposed method uses an edge federated learning architecture to optimize task allocation and execution. Unlike in traditional edge networks where requests from fixed nodes are handled by nearby edge devices or remote clouds, the proposed model uses an intelligent broker within the federation to assess member edge cloudlets' parameters, such as resources and hop count, to make optimal decisions for task offloading. This approach enhances performance and privacy by placing tasks in closer proximity to the user. DenseNet is used for model training, with a depth of 60 and 357,482 parameters. This resource-aware distributed approach optimizes computing resource utilization within the edge-federated learning architecture. Results The experimental results demonstrate significant improvements in various performance metrics. The proposed method reduces training time by 53.1%, optimizes CPU and memory utilization by 17.5% and 33.6%, and maintains accurate COVID-19 detection capabilities without compromising the F1 score, demonstrating the efficiency and effectiveness of the lightweight convolutional neural network-based edge federated learning architecture. Conclusion Existing studies predominantly concentrate on either privacy and accuracy or load balancing and energy optimization, with limited emphasis on training time. The proposed approach offers a comprehensive performance-centric solution that simultaneously addresses privacy, load balancing, and energy optimization while reducing training time, providing a more holistic and balanced solution for optimal system performance. metadata Alvi, Sohaib Bin Khalid and Nayyer, Muhammad Ziad and Jamal, Muhammad Hasan and Raza, Imran and de la Torre Diez, Isabel and Rodríguez Velasco, Carmen Lilí and Breñosa, Jose and Ashraf, Imran mail UNSPECIFIED, UNSPECIFIED, UNSPECIFIED, UNSPECIFIED, UNSPECIFIED, carmen.rodriguez@uneatlantico.es, josemanuel.brenosa@uneatlantico.es, UNSPECIFIED (2023) A lightweight deep learning approach for COVID-19 detection using X-ray images with edge federation. DIGITAL HEALTH, 9. ISSN 2055-2076

[img] Text
alvi-et-al-2023-a-lightweight-deep-learning-approach-for-covid-19-detection-using-x-ray-images-with-edge-federation.pdf
Available under License Creative Commons Attribution Non-commercial.

Download (1MB)

Abstract

Objective This study aims to develop a lightweight convolutional neural network-based edge federated learning architecture for COVID-19 detection using X-ray images, aiming to minimize computational cost, latency, and bandwidth requirements while preserving patient privacy. Method The proposed method uses an edge federated learning architecture to optimize task allocation and execution. Unlike in traditional edge networks where requests from fixed nodes are handled by nearby edge devices or remote clouds, the proposed model uses an intelligent broker within the federation to assess member edge cloudlets' parameters, such as resources and hop count, to make optimal decisions for task offloading. This approach enhances performance and privacy by placing tasks in closer proximity to the user. DenseNet is used for model training, with a depth of 60 and 357,482 parameters. This resource-aware distributed approach optimizes computing resource utilization within the edge-federated learning architecture. Results The experimental results demonstrate significant improvements in various performance metrics. The proposed method reduces training time by 53.1%, optimizes CPU and memory utilization by 17.5% and 33.6%, and maintains accurate COVID-19 detection capabilities without compromising the F1 score, demonstrating the efficiency and effectiveness of the lightweight convolutional neural network-based edge federated learning architecture. Conclusion Existing studies predominantly concentrate on either privacy and accuracy or load balancing and energy optimization, with limited emphasis on training time. The proposed approach offers a comprehensive performance-centric solution that simultaneously addresses privacy, load balancing, and energy optimization while reducing training time, providing a more holistic and balanced solution for optimal system performance.

Item Type: Article
Uncontrolled Keywords: Public health, federated learning, edge computing, deep learning
Subjects: Subjects > Biomedicine
Subjects > Engineering
Divisions: Europe University of Atlantic > Research > Scientific Production
Fundación Universitaria Internacional de Colombia > Research > Scientific Production
Ibero-american International University > Research > Articles and books
Ibero-american International University > Research > Scientific Production
Universidad Internacional do Cuanza > Research > Scientific Production
Date Deposited: 17 Oct 2023 23:30
Last Modified: 17 Oct 2023 23:30
URI: https://repositorio.unini.edu.mx/id/eprint/9229

Actions (login required)

View Item View Item

<a class="ep_document_link" href="/10290/1/Influence%20of%20E-learning%20training%20on%20the%20acquisition%20of%20competences%20in%20basketball%20coaches%20in%20Cantabria.pdf"><img class="ep_doc_icon" alt="[img]" src="/style/images/fileicons/text.png" border="0"/></a>

en

open

Influence of E-learning training on the acquisition of competences in basketball coaches in Cantabria

The main aim of this study was to analyse the influence of e-learning training on the acquisition of competences in basketball coaches in Cantabria. The current landscape of basketball coach training shows an increasing demand for innovative training models and emerging pedagogies, including e-learning-based methodologies. The study sample consisted of fifty students from these courses, all above 16 years of age (36 males, 14 females). Among them, 16% resided outside the autonomous community of Cantabria, 10% resided more than 50 km from the city of Santander, 36% between 10 and 50 km, 14% less than 10 km, and 24% resided within Santander city. Data were collected through a Google Forms survey distributed by the Cantabrian Basketball Federation to training course students. Participation was voluntary and anonymous. The survey, consisting of 56 questions, was validated by two sports and health doctors and two senior basketball coaches. The collected data were processed and analysed using Microsoft® Excel version 16.74, and the results were expressed in percentages. The analysis revealed that 24.60% of the students trained through the e-learning methodology considered themselves fully qualified as basketball coaches, contrasting with 10.98% of those trained via traditional face-to-face methodology. The results of the study provide insights into important characteristics that can be adjusted and improved within the investigated educational process. Moreover, the study concludes that e-learning training effectively qualifies basketball coaches in Cantabria.

Producción Científica

Josep Alemany Iturriaga mail josep.alemany@uneatlantico.es, Álvaro Velarde-Sotres mail alvaro.velarde@uneatlantico.es, Javier Jorge mail , Kamil Giglio mail ,

Alemany Iturriaga

<a class="ep_document_link" href="/11941/1/healthcare-12-00942.pdf"><img class="ep_doc_icon" alt="[img]" src="/style/images/fileicons/text.png" border="0"/></a>

en

open

Risk Factors for Eating Disorders in University Students: The RUNEAT Study

The purpose of the study is to assess the risk of developing general eating disorders (ED), anorexia nervosa (AN), and bulimia nervosa (BN), as well as to examine the effects of gender, academic year, place of residence, faculty, and diet quality on that risk. Over two academic years, 129 first- and fourth-year Uneatlántico students were included in an observational descriptive study. The self-administered tests SCOFF, EAT-26, and BITE were used to determine the participants’ risk of developing ED. The degree of adherence to the Mediterranean diet (MD) was used to evaluate the quality of the diet. Data were collected at the beginning (T1) and at the end (T2) of the academic year. The main results were that at T1, 34.9% of participants were at risk of developing general ED, AN 3.9%, and BN 16.3%. At T2, these percentages were 37.2%, 14.7%, and 8.5%, respectively. At T2, the frequency of general ED in the female group was 2.5 times higher (OR: 2.55, 95% CI: 1.22–5.32, p = 0.012). The low-moderate adherence to the MD students’ group was 0.92 times less frequent than general ED at T2 (OR: 0.921, 95%CI: 0.385–2.20, p < 0.001). The most significant risk factor for developing ED is being a female in the first year of university. Moreover, it appears that the likelihood of developing ED generally increases during the academic year.

Producción Científica

Imanol Eguren García mail imanol.eguren@uneatlantico.es, Sandra Sumalla Cano mail sandra.sumalla@uneatlantico.es, Sandra Conde González mail , Anna Vila-Martí mail , Mercedes Briones Urbano mail mercedes.briones@uneatlantico.es, Raquel Martínez Díaz mail raquel.martinez@uneatlantico.es, Iñaki Elío Pascual mail inaki.elio@uneatlantico.es,

Eguren García

<a class="ep_document_link" href="/11642/1/s41598-024-57547-4.pdf"><img class="ep_doc_icon" alt="[img]" src="/style/images/fileicons/text.png" border="0"/></a>

en

open

Adaptive neighborhood rough set model for hybrid data processing: a case study on Parkinson’s disease behavioral analysis

Extracting knowledge from hybrid data, comprising both categorical and numerical data, poses significant challenges due to the inherent difficulty in preserving information and practical meanings during the conversion process. To address this challenge, hybrid data processing methods, combining complementary rough sets, have emerged as a promising approach for handling uncertainty. However, selecting an appropriate model and effectively utilizing it in data mining requires a thorough qualitative and quantitative comparison of existing hybrid data processing models. This research aims to contribute to the analysis of hybrid data processing models based on neighborhood rough sets by investigating the inherent relationships among these models. We propose a generic neighborhood rough set-based hybrid model specifically designed for processing hybrid data, thereby enhancing the efficacy of the data mining process without resorting to discretization and avoiding information loss or practical meaning degradation in datasets. The proposed scheme dynamically adapts the threshold value for the neighborhood approximation space according to the characteristics of the given datasets, ensuring optimal performance without sacrificing accuracy. To evaluate the effectiveness of the proposed scheme, we develop a testbed tailored for Parkinson’s patients, a domain where hybrid data processing is particularly relevant. The experimental results demonstrate that the proposed scheme consistently outperforms existing schemes in adaptively handling both numerical and categorical data, achieving an impressive accuracy of 95% on the Parkinson’s dataset. Overall, this research contributes to advancing hybrid data processing techniques by providing a robust and adaptive solution that addresses the challenges associated with handling hybrid data, particularly in the context of Parkinson’s disease analysis.

Producción Científica

Imran Raza mail , Muhammad Hasan Jamal mail , Rizwan Qureshi mail , Abdul Karim Shahid mail , Angel Olider Rojas Vistorte mail angel.rojas@uneatlantico.es, Md Abdus Samad mail , Imran Ashraf mail ,

Raza

<a href="/12103/1/jcm-13-02514.pdf" class="ep_document_link"><img class="ep_doc_icon" alt="[img]" src="/style/images/fileicons/text.png" border="0"/></a>

en

open

A Prospective Observational Study of Frailty in Geriatric Revitalization Aimed at Community-Dwelling Elderly

(1) Background: The increasing life expectancy brings an increase in geriatric syndromes, specifically frailty. The literature shows that exercise is a key to preventing, or even reversing, frailty in community-dwelling populations. The main objective is to demonstrate how an intervention based on multicomponent exercise produces an improvement in frailty and pre-frailty in a community-dwelling population. (2) Methods: a prospective observational study of a multicomponent exercise program for geriatric revitalization with people aged over 65 holding Barthel Index scores equal to, or beyond, 90. The program was developed over 30 weeks, three times a week, in sessions lasting 45–50 min each. Frailty levels were registered by the Short Physical Performance Battery, FRAIL Questionnaire Screening Tool, and Timed “Up & Go” at the beginning of the program, 30 weeks later (at the end of the program), and following 13 weeks without training; (3) Results: 360 participants completed the program; a greater risk of frailty was found before the program started among older women living in urban areas, with a more elevated fat percentage, more baseline pathologies, and wider baseline medication use. Furthermore, heterogeneous results were observed both in training periods and in periods without physical activity. However, they are consistent over time and show improvement after training. They show a good correlation between TUG and SPPB; (4) Conclusions: A thirty-week multicomponent exercise program improves frailty and pre-frailty status in a community-dwelling population with no functional decline. Nevertheless, a lack of homogeneity is evident among the various tools used for measuring frailty over training periods and inactivity periods.

Producción Científica

Almudena Morales-Sánchez mail , José Ignacio Calvo Arenillas mail , María José Gutiérrez Palmero mail , José L. Martín-Conty mail , Begoña Polonio-López mail , Luis Alonso Dzul Lopez mail luis.dzul@unini.edu.mx, Laura Mordillo-Mateos mail , Juan José Bernal-Jiménez mail , Rosa Conty-Serrano mail , Francisca Torres-Falguera mail , Alfonso Martínez Cano mail , Carlos Durantez-Fernández mail ,

Morales-Sánchez

<a href="/12370/1/dwivedi-et-al-2024-exploring-microstructural-interfacial-mechanical-and-wear-properties-of-alsi7mg0-3-composites-with.pdf" class="ep_document_link"><img class="ep_doc_icon" alt="[img]" src="/style/images/fileicons/text.png" border="0"/></a>

en

open

Exploring Microstructural, Interfacial, Mechanical, and Wear Properties of AlSi7Mg0.3 Composites with TiMOVWCr High-Entropy Alloy Powder

This study explored the impact of varying weight percentages of TiMoVWCr high-entropy alloy (HEA) powder addition on A356 composites produced using friction stir processing (FSP). Unlike previous research that often focused on singular aspects, such as mechanical properties, or microstructural analysis, this investigation systematically examined the multifaceted performance of A356 composites by comprehensively assessing the microstructure, interfacial bonding strength, mechanical properties, and wear behavior. The study identified a uniform distribution of TiMoVWCr HEA powder in the composition A356/2%Ti2%Mo2%V2%W2%Cr, highlighting the effectiveness of the FSP technique in achieving homogeneous dispersion. Strong bonding between the reinforcement and matrix material was observed in the same composition, indicating favorable interfacial characteristics. Mechanical properties, including tensile strength and hardness, were evaluated for various compositions, demonstrating significant improvements across the board. The addition of 2%Ti2%Mo2%V2%W2%Cr powder enhanced the tensile strength by 36.39%, while hardness improved by 62.71%. Similarly, wear resistance showed notable enhancements ranging from 35.56 to 48.89% for different compositions. Microstructural analysis revealed approximately 1640.59 grains per square inch for the A356/2%Ti2%Mo2%V2%W2%Cr processed composite at 500 magnifications. In reinforcing Al composites with Ti, Mo, V, W, and Cr high-entropy alloy (HEA) particles, each element imparted distinct benefits. Titanium (Ti) enhanced strength and wear resistance, molybdenum (Mo) contributed to improved hardness, vanadium (V) promoted hardenability, tungsten (W) enhanced wear resistance, and chromium (Cr) provided wear resistance and hardness. Anticipating the potential applications of the developed composite, the study suggests its suitability for the aerospace sector, particularly in casting lightweight yet high-strength parts such as aircraft components, engine components, and structural components, underlining the significance of the investigated TiMoVWCr HEA powder-modified A356 composites.

Artículos y libros

Shashi Prakash Dwivedi mail , Shubham Sharma mail , Changhe Li mail , Yanbin Zhang mail , Rajesh Singh mail , Abhinav Kumar mail , Fuad A. Awwad mail , M. Ijaz Khan mail , Emad A. A. Ismail mail ,

Dwivedi