[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
5.3
2.6
Journals
Electronics
Special Issues
Recent Trends in Artificial Learning and Data Processing for Biomedical...
share announcement

Recent Trends in Artificial Learning and Data Processing for Biomedical Engineering

A special issue of Electronics (ISSN 2079-9292). This special issue belongs to the section "Bioelectronics".

Deadline for manuscript submissions: 15 April 2025 | Viewed by 10489

Special Issue Editors

Prof. Dr. Larbi Boubchir

E-Mail Website
Guest Editor
LIASD Research Lab., University of Paris 8, 2 Rue de la Liberté, 93526 Saint-Denis, France
Interests: biomedical signal processing; EEG; image processing; machine learning; brain–computer interface; biometrics
Special Issues, Collections and Topics in MDPI journals
Prof. Dr. Boubaker Daachi

E-Mail Website
Guest Editor
LIASD Research Lab. – University of Paris 8, 2 Rue de la Liberté, 93526 Saint-Denis, France
Interests: robotics; soft computing; BCI; WSN; biometrics
Special Issues, Collections and Topics in MDPI journals

Special Issue Information

Dear Colleagues,

This Special Issue aims to solicit original research papers focusing on novel solutions to challenging problems in biomedical engineering using artificial learning and advanced data processing algorithms and methods.

We are inviting original research works covering novel theories, innovative methods, and meaningful applications that can potentially lead to significant advances in the biomedical engineering field.

The main topics of interest include but are not limited to:

  • Biomedical signal processing;
  • Medical and biological imaging;
  • Pattern recognition algorithms and methods;
  • Artificial learning algorithms and methods (e.g., machine learning, deep learning, statistical learning);
  • Applications of artificial intelligence in biomedical engineering;
  • Healthcare applications (e.g., detection, diagnostic, therapeutic, e-health, m-health);
  • Healthcare Internet of Things;
  • Smart Healthcare;
  • Decision support systems in biomedical engineering;
  • Neural engineering;
  • Clinical engineering;
  • Rehabilitation engineering;
  • Biological engineering;
  • Biomedical sensors and devices;
  • Biomedical wearable technology;
  • Related applications.

Prof. Dr. Larbi Boubchir
Prof. Dr. Boubaker Daachi 
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Electronics is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

  • image processing
  • pattern recognition
  • artificial intelligence
  • machine learning
  • feature engineering
  • biomedical engineering
  • healthcare

Benefits of Publishing in a Special Issue

  • Ease of navigation: Grouping papers by topic helps scholars navigate broad scope journals more efficiently.
  • Greater discoverability: Special Issues support the reach and impact of scientific research. Articles in Special Issues are more discoverable and cited more frequently.
  • Expansion of research network: Special Issues facilitate connections among authors, fostering scientific collaborations.
  • External promotion: Articles in Special Issues are often promoted through the journal's social media, increasing their visibility.
  • e-Book format: Special Issues with more than 10 articles can be published as dedicated e-books, ensuring wide and rapid dissemination.

Further information on MDPI's Special Issue policies can be found here.

Published Papers (6 papers)

Download All Papers
Order results
Result details
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:

Research

Jump to: Other

20 pages, 3455 KiB  
Article
Improved EfficientNet Architecture for Multi-Grade Brain Tumor Detection
by Ahmad Ishaq, Fath U Min Ullah, Prince Hamandawana, Da-Jung Cho and Tae-Sun Chung
Electronics 2025, 14(4), 710; https://doi.org/10.3390/electronics14040710 - 12 Feb 2025
Viewed by 529
Abstract
Accurate detection and diagnosis of brain tumors at early stages is significant for effective treatment. While numerous methods have been developed for tumor detection and classification, several rely on traditional techniques, often resulting in suboptimal performance. In contrast, AI-based deep learning techniques have [...] Read more.
Accurate detection and diagnosis of brain tumors at early stages is significant for effective treatment. While numerous methods have been developed for tumor detection and classification, several rely on traditional techniques, often resulting in suboptimal performance. In contrast, AI-based deep learning techniques have shown promising results, consistently achieving high accuracy across various tumor types while maintaining model interpretability. Inspired by these advancements, this paper introduces an improved variant of EfficientNet for multi-grade brain tumor detection and classification, addressing the gap between performance and explainability. Our approach extends the capabilities of EfficientNet to classify four tumor types: glioma, meningioma, pituitary tumor, and non-tumor. For enhanced explainability, we incorporate gradient-weighted class activation mapping (Grad-CAM) to improve model interpretability. The input MRI images undergo data augmentation before being passed through the feature extraction phase, where the underlying tumor patterns are learned. Our model achieves an average accuracy of 98.6%, surpassing other state-of-the-art methods on standard datasets while maintaining a substantially reduced parameter count. Furthermore, the explainable AI (XAI) analysis demonstrates the model’s ability to focus on relevant tumor regions, enhancing its interpretability. This accurate and interpretable model for brain tumor classification has the potential to significantly aid clinical decision-making in neuro-oncology. Full article
(This article belongs to the Special Issue Recent Trends in Artificial Learning and Data Processing for Biomedical Engineering)
►▼ Show Figures

Figure 1

Figure 1
<p>Overview of the proposed model framework where the pre-processed MRI input data, obtained through data splitting, filtering, augmentation, and resizing, is fed into the feature extraction network. Based on the features, the trained model classifies the input images as glioma, meningioma, pituitary tumor, or non-tumor categories. Input image sample taken from the BT-3264 dataset [<a href="#B17-electronics-14-00710" class="html-bibr">17</a>].</p> Full article ">Figure 2
<p>Tools applied to augment the MRI images. Input image sample taken from the BT-3264 dataset [<a href="#B17-electronics-14-00710" class="html-bibr">17</a>].</p> Full article ">Figure 3
<p>Adopted transfer learning method.</p> Full article ">Figure 4
<p>Accuracy and loss comparison between baseline and customized approach using small dataset BT_3264.</p> Full article ">Figure 5
<p>Accuracy and loss comparison between baseline and customized mode using larger dataset BT_7023.</p> Full article ">Figure 6
<p>Confusion matrix of BT_7023 with baseline and customized model.</p> Full article ">Figure 7
<p>Accuracy and loss comparison between baseline and customized mode using Figshare dataset.</p> Full article ">Figure 8
<p>Visualization of different tumor types, showing the heat-maps for the original and augmented images (with labels). Each row displays the original MRI scan, the corresponding heatmap visualization, the augmented MRI scan with increased contrast and brightness, and the heatmap of the augmented image. Input image sample taken from the BT-3264 dataset [<a href="#B17-electronics-14-00710" class="html-bibr">17</a>].</p> Full article ">
19 pages, 688 KiB  
Article
Advancing Pulmonary Nodule Detection with ARSGNet: EfficientNet and Transformer Synergy
by Maroua Oumlaz, Yassine Oumlaz, Aziz Oukaira, Amrou Zyad Benelhaouare and Ahmed Lakhssassi
Electronics 2024, 13(22), 4369; https://doi.org/10.3390/electronics13224369 - 7 Nov 2024
Viewed by 994
Abstract
Lung cancer, the leading cause of cancer-related deaths globally, presents significant challenges in early detection and diagnosis. The effective analysis of pulmonary medical imaging, particularly computed tomography (CT) scans, is critical in this endeavor. Traditional diagnostic methods, which are manual and time-intensive, underscore [...] Read more.
Lung cancer, the leading cause of cancer-related deaths globally, presents significant challenges in early detection and diagnosis. The effective analysis of pulmonary medical imaging, particularly computed tomography (CT) scans, is critical in this endeavor. Traditional diagnostic methods, which are manual and time-intensive, underscore the need for innovative, efficient, and accurate detection approaches. To address this need, we introduce the Adaptive Range Slice Grouping Network (ARSGNet), a novel deep learning framework that enhances early lung cancer diagnosis through advanced segmentation and classification techniques in CT imaging. ARSGNet synergistically integrates the strengths of EfficientNet and Transformer architectures, leveraging their superior feature extraction and contextual processing capabilities. This hybrid model proficiently handles the complexities of 3D CT images, ensuring precise and reliable lung nodule detection. The algorithm processes CT scans using short slice grouping (SSG) and long slice grouping (LSG) techniques to extract critical features from each slice, culminating in the generation of nodule probabilities and the identification of potential nodular regions. Incorporating shapley additive explanations (SHAP) analysis further enhances model interpretability by highlighting the contributory features. Our extensive experimentation demonstrated a significant improvement in diagnostic accuracy, with training accuracy increasing from 0.9126 to 0.9817. This advancement not only reflects the model’s efficient learning curve but also its high proficiency in accurately classifying a majority of training samples. Given its high accuracy, interpretability, and consistent reduction in training loss, ARSGNet holds substantial potential as a groundbreaking tool for early lung cancer detection and diagnosis. Full article
(This article belongs to the Special Issue Recent Trends in Artificial Learning and Data Processing for Biomedical Engineering)
►▼ Show Figures

Figure 1

Figure 1
<p>Pixel intensity distribution of the training dataset after preprocessing.</p> Full article ">Figure 2
<p>Pixel intensity distribution of the testing dataset after preprocessing.</p> Full article ">Figure 3
<p>Workflow of ARSGNet for Pulmonary nodule detection.</p> Full article ">Figure 4
<p>Loss, Accuracy, Dice Score, and IOU score for ARGSnet in training vs. validation.</p> Full article ">
24 pages, 6207 KiB  
Article
Dynamic Partitioning of Graphs Based on Multivariate Blood Glucose Data—A Graph Neural Network Model for Diabetes Prediction
by Jianjun Li, Xiaozhe Jiang and Kaiyue Wang
Electronics 2024, 13(18), 3727; https://doi.org/10.3390/electronics13183727 - 20 Sep 2024
Cited by 1 | Viewed by 1166
Abstract
Postprandial Hyperglycemia (PPHG) persistently threatens patients’ health. Therefore, accurate diabetes prediction is crucial for effective blood glucose management. Most current methods primarily focus on analyzing univariate blood glucose data using traditional neural networks, neglecting the importance of spatiotemporal modeling of multivariate data at [...] Read more.
Postprandial Hyperglycemia (PPHG) persistently threatens patients’ health. Therefore, accurate diabetes prediction is crucial for effective blood glucose management. Most current methods primarily focus on analyzing univariate blood glucose data using traditional neural networks, neglecting the importance of spatiotemporal modeling of multivariate data at the node and subgraph levels. This study aimed to evaluate the accuracy of using deep learning (DL) techniques to predict diabetes based on multivariable blood glucose data, aiming to improve resource allocation and decision-making in healthcare. We introduce a Nonlinear Aggregated Graph Neural Network (NLAGNN) that utilizes continuous multivariate historical blood glucose data from multiple patients to predict blood glucose levels over time, addressing the challenge of accurately extracting strong and weak correlation features. We preliminarily propose a Nonlinear Fourier Graph Neural Operator (NFGO) for nonlinear node representation, which effectively reduces meaningless noise. Additionally, a dynamic partitioning of graphs is introduced, which divides the a hypergraph into distinct subgraphs, enabling the further processing of strongly correlated features at the node and subgraph levels, ultimately obtaining the final prediction through layer aggregation. Extensive experiments on three datasets show that our proposed method achieves competitive results compared to existing advanced methods. Full article
(This article belongs to the Special Issue Recent Trends in Artificial Learning and Data Processing for Biomedical Engineering)
►▼ Show Figures

Figure 1

Figure 1
<p>The overall framework of the NLAGNN.</p> Full article ">Figure 2
<p>Clarke Error grid analysis results graph. (<b>a</b>) D1NAMO-1 univariate glucose prediction error. (<b>b</b>) D1NAMO-1 univariate glucose prediction error. (<b>c</b>) D1NAMO-2 multivariate glucose prediction error. (<b>d</b>) D1NAMO-2 multivariate glucose prediction error. (<b>e</b>) DirecNet univariate glucose prediction error. (<b>f</b>) DirecNet univariate glucose prediction error. (<b>g</b>) T1DExchange univariate glucose prediction error. (<b>h</b>) T1DExchange univariate glucose prediction error.</p> Full article ">Figure 3
<p>Forecast Curves at different prediction horizons in various datasets. (<b>a</b>) D1NAMO-1 univariate glucose prediction curve. (<b>b</b>) D1NAMO-1 univariate glucose prediction curve. (<b>c</b>) D1NAMO-2 multivariate glucose prediction curve. (<b>d</b>) D1NAMO-2 multivariate glucose prediction curve. (<b>e</b>) DirecNet univariate glucose prediction curve. (<b>f</b>) DirecNet univariate glucose prediction curve. (<b>g</b>) T1DExchange univariate glucose prediction curve. (<b>h</b>) T1DExchange univariate glucose prediction curve.</p> Full article ">
22 pages, 577 KiB  
Article
Efficient Human Activity Recognition on Wearable Devices Using Knowledge Distillation Techniques
by Paulo H. N. Gonçalves, Hendrio Bragança and Eduardo Souto
Electronics 2024, 13(18), 3612; https://doi.org/10.3390/electronics13183612 - 11 Sep 2024
Viewed by 1416
Abstract
Mobile and wearable devices have revolutionized the field of continuous user activity monitoring. However, analyzing the vast and intricate data captured by the sensors of these devices poses significant challenges. Deep neural networks have shown remarkable accuracy in Human Activity Recognition (HAR), but [...] Read more.
Mobile and wearable devices have revolutionized the field of continuous user activity monitoring. However, analyzing the vast and intricate data captured by the sensors of these devices poses significant challenges. Deep neural networks have shown remarkable accuracy in Human Activity Recognition (HAR), but their application on mobile and wearable devices is constrained by limited computational resources. To address this limitation, we propose a novel method called Knowledge Distillation for Human Activity Recognition (KD-HAR) that leverages the knowledge distillation technique to compress deep neural network models for HAR using inertial sensor data. Our approach transfers the acquired knowledge from high-complexity teacher models (state-of-the-art models) to student models with reduced complexity. This compression strategy allows us to maintain performance while keeping computational costs low. To assess the compression capabilities of our approach, we evaluate it using two popular databases (UCI-HAR and WISDM) comprising inertial sensor data from smartphones. Our results demonstrate that our method achieves competitive accuracy, even at compression rates ranging from 18 to 42 times the number of parameters compared to the original teacher model. Full article
(This article belongs to the Special Issue Recent Trends in Artificial Learning and Data Processing for Biomedical Engineering)
►▼ Show Figures

Figure 1

Figure 1
<p>The knowledge distillation architecture transfers knowledge from a complex teacher model to a simpler student model.</p> Full article ">Figure 2
<p>Overview of the proposed KD-HAR method for compressing neural networks based on the knowledge distillation technique applied to human activity recognition models.</p> Full article ">Figure 3
<p>The architecture used to train the teacher model.</p> Full article ">Figure 4
<p>The network architecture used to generate the student model.</p> Full article ">Figure 5
<p>Knowledge distillation process. The final loss function is obtained taking into account a loss function on the class probabilities of teachers and students smoothed by the temperature <span class="html-italic">T</span>, which is multiplied by a factor <math display="inline"><semantics> <mi>β</mi> </semantics></math> and added to the function of student loss without smoothing multiplied by <math display="inline"><semantics> <mi>α</mi> </semantics></math>.</p> Full article ">Figure 6
<p>Example of total number of parameters from Keras Framework.</p> Full article ">Figure 7
<p>UCI dataset F1-score values for the best <math display="inline"><semantics> <mrow> <mi>K</mi> <mi>D</mi> </mrow> </semantics></math>, <span class="html-italic">S</span> and <span class="html-italic">P</span> models.</p> Full article ">Figure 8
<p>UCI confusion matrix of the <span class="html-italic">P</span>, best <math display="inline"><semantics> <mrow> <mi>K</mi> <mi>D</mi> </mrow> </semantics></math> and <span class="html-italic">S</span> models.</p> Full article ">Figure 9
<p>WISDM dataset F1-score values for the <span class="html-italic">P</span>, <span class="html-italic">S</span> and best <math display="inline"><semantics> <mrow> <mi>K</mi> <mi>D</mi> </mrow> </semantics></math> models. In addiction, F1-score values for Peppas et al. [<a href="#B28-electronics-13-03612" class="html-bibr">28</a>] and Ignatov et al. [<a href="#B11-electronics-13-03612" class="html-bibr">11</a>].</p> Full article ">Figure 10
<p>WISDM confusion matrix of the <span class="html-italic">P</span>, best <math display="inline"><semantics> <mrow> <mi>K</mi> <mi>D</mi> </mrow> </semantics></math>, and <span class="html-italic">S</span> models.</p> Full article ">
15 pages, 5290 KiB  
Article
An Empirical Mode Decomposition-Based Method to Identify Topologically Associated Domains from Chromatin Interactions
by Xuemin Zhao, Ran Duan and Shaowen Yao
Electronics 2023, 12(19), 4154; https://doi.org/10.3390/electronics12194154 - 6 Oct 2023
Cited by 1 | Viewed by 1457
Abstract
Topologically associated domains (TADs) represent essential units constituting chromatin’s intricate three-dimensional spatial organization. TADs are stably present across cell types and species, and their influence on vital biological processes, such as gene expression, DNA replication, and chromosomal translocation, underscores their significance. Accordingly, the [...] Read more.
Topologically associated domains (TADs) represent essential units constituting chromatin’s intricate three-dimensional spatial organization. TADs are stably present across cell types and species, and their influence on vital biological processes, such as gene expression, DNA replication, and chromosomal translocation, underscores their significance. Accordingly, the identification of TADs within the Hi-C interaction matrix is a key point in three-dimensional genomics. TADs manifest as contiguous blocks along the diagonal of the Hi-C interaction matrix, which are characterized by dense interactions within blocks and sparse interactions between blocks. An optimization method is proposed to enhance Hi-C interaction matrix data using the empirical mode decomposition method, which requires no prior knowledge and adaptively decomposes Hi-C data into a sum of multiple eigenmodal functions via exploiting the inherent characteristics of variations in the input Hi-C data. We identify TADs within the optimized data and compared the results with five commonly used TAD detection methods, namely the Directionality Index (DI), Interaction Isolation (IS), HiCKey, HiCDB, and TopDom. The results demonstrate the universality and efficiency of the proposed method, highlighting its potential as a valuable tool in TAD identification. Full article
(This article belongs to the Special Issue Recent Trends in Artificial Learning and Data Processing for Biomedical Engineering)
►▼ Show Figures

Figure 1

Figure 1
<p>Heat map of the Hi-C interaction matrix at various treatment stages: (<b>A</b>) raw Hi-C interaction matrix; (<b>B</b>) ICE normalized interaction matrix; (<b>C</b>) fusion structure data interaction matrix. (Rao2014-GM12878-DpnII-allreps-filtered-50 kb.).</p> Full article ">Figure 2
<p>Framework diagram for the EMTAD methodology.</p> Full article ">Figure 3
<p>A comparative analysis of six different methods for the identification of TADs: (<b>A</b>) Evaluation of the total number of TADs and their distribution patterns. (<b>B</b>) Comparative evaluation of TAD boundary scores specifically within chromosome 2. (<b>C</b>) Investigation of the DI and IS coefficients associated with TAD boundaries. (<b>D</b>) Comparison of profile coefficients associated with TADs. (<b>E</b>) Evaluation of interaction coefficients associated with TADs. (<b>F</b>) Consistency comparison of identification results.</p> Full article ">Figure 4
<p>Results of simulated data analysis.</p> Full article ">Figure 5
<p>Comparison the positions of TADs identified using data at different resolutions: (<b>A</b>) consistency comparison of the six methods at 25 Kb vs. 50 Kb resolution, 25 Kb vs. 100 Kb resolution, and 50 Kb vs. 100 Kb resolution; (<b>B</b>) consistency comparison at different sampling depths; (<b>C</b>) consistency comparison at different resolutions and different sampling depths.</p> Full article ">Figure 6
<p>Comparison of results obtained using different methods for evaluating enrichment of TAD boundary transcription factors and TAD regional histone modifications: (<b>A</b>) comparison of CTCF enrichment analysis; (<b>B</b>) comparison of RAD21 enrichment analysis; (<b>C</b>) comparison of SMC3 enrichment analysis; (<b>D</b>) comparison of histone enrichment analysis of TAD regions.</p> Full article ">

Other

Jump to: Research

29 pages, 1905 KiB  
Systematic Review
Health Risk Assessment Using Machine Learning: Systematic Review
by Stanley Ebhohimhen Abhadiomhen, Emmanuel Onyekachukwu Nzeakor and Kiemute Oyibo
Electronics 2024, 13(22), 4405; https://doi.org/10.3390/electronics13224405 - 11 Nov 2024
Viewed by 2991
Abstract
According to the World Health Organization, chronic illnesses account for over 70% of deaths globally, underscoring the need for effective health risk assessment (HRA). While machine learning (ML) has shown potential in enhancing HRA, no systematic review has explored its application in general [...] Read more.
According to the World Health Organization, chronic illnesses account for over 70% of deaths globally, underscoring the need for effective health risk assessment (HRA). While machine learning (ML) has shown potential in enhancing HRA, no systematic review has explored its application in general health risk assessments. Existing reviews typically focus on specific conditions. This paper reviews published articles that utilize ML for HRA, and it aims to identify the model development methods. A systematic review following Tranfield et al.’s three-stage approach was conducted, and it adhered to the PRISMA protocol. The literature was sourced from five databases, including PubMed. Of the included articles, 42% (11/26) addressed general health risks. Secondary data sources were most common (14/26, 53.85%), while primary data were used in eleven studies, with nine (81.81%) using data from a specific population. Random forest was the most popular algorithm, which was used in nine studies (34.62%). Notably, twelve studies implemented multiple algorithms, while seven studies incorporated model interpretability techniques. Although these studies have shown promise in addressing digital health inequities, more research is needed to include diverse sample populations, particularly from underserved communities, to enhance the generalizability of existing models. Furthermore, model interpretability should be prioritized to ensure transparent, trustworthy, and broadly applicable healthcare solutions. Full article
(This article belongs to the Special Issue Recent Trends in Artificial Learning and Data Processing for Biomedical Engineering)
►▼ Show Figures

Figure 1

Figure 1
<p>A typical machine learning workflow for health risk assessment, encompassing data collection, data annotation, data pre-processing to clean and prepare the data, and model training.</p> Full article ">Figure 2
<p>PRISMA flowchart for the screening and inclusion of articles in the systematic review. WOS: Web of Science. TA: Title, abstract screening.</p> Full article ">Figure 3
<p>Publication distribution by year: (<b>a</b>) Total publications; (<b>b</b>) Journal publications; (<b>c</b>) Conference publications; (<b>d</b>) The proportion of journal and conference publications obtained by dividing the respective counts by the total counts.</p> Full article ">Figure 4
<p>Stacked bar chart illustrating the distribution of bias concerns across various risk domains.</p> Full article ">
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-6
Back to TopTop