Journal of Medical Signals & Sensors最新文献

{"title":"DenseUNet for Breast Cancer Segmentation in Histopathological Images.","authors":"Habib Rasi, Hossein Ebrahimnezhad, Mohammad Hossein Sedaaghi","doi":"10.4103/jmss.jmss_83_25","DOIUrl":"10.4103/jmss.jmss_83_25","url":null,"abstract":"<p><strong>Background: </strong>Breast cancer remains one of the leading causes of mortality among women worldwide, highlighting the urgent need for accurate and efficient diagnostic tools. Histopathological image analysis plays a critical role in diagnosis by enabling cellularlevel tissue examination. However, manual assessment is often timeconsuming, subjective, and prone to variability, driving the need for automated solutions.</p><p><strong>Methods: </strong>This study proposes DenseUNet169, a deep learning framework that integrates established DenseNet and UNet for systematic evaluation in breast cancer segmentation. The model improves gradient flow, facilitates efficient training, and enhances segmentation accuracy by focusing on visually salient regions within histopathological images. Performance is evaluated using key metrics, including Intersection over Union, Dice Coefficient, Pixel Accuracy, Modified Hausdorff Distance, Surface Dice Overlap, Log Loss, and Jaccard Index.</p><p><strong>Results: </strong>Experiments conducted on a publicly available breast cancer segmentation dataset demonstrate competitive performance compared to alternative backbone configurations.</p><p><strong>Conclusion: </strong>Given the limited dataset size, the results should be interpreted as a proof of concept evaluation of the proposed architectural integration rather than evidence of largescale clinical deployment readiness.</p>","PeriodicalId":37680,"journal":{"name":"Journal of Medical Signals & Sensors","volume":"16 ","pages":"20"},"PeriodicalIF":1.8,"publicationDate":"2026-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13557525/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148921243","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Implementing a Version Control Platform as an Interactive Educational Framework for Bioinformatics and Artificial Intelligence in Medicine.","authors":"Alireza Ani, Ahmad Vaez","doi":"10.4103/jmss.jmss_2_26","DOIUrl":"10.4103/jmss.jmss_2_26","url":null,"abstract":"<p><strong>Background: </strong>Complex fields such as bioinformatics and artificial intelligence (AI) in medicine depend on big data and advanced computational tools. Consequently, they require structured platforms that enable collaborative development and academic capabilities. This report aims to demonstrate the pedagogical value of a version control system in an academic setting by deploying a dedicated local server for teaching AI in the field of medical sciences and developing bioinformatics software tools.</p><p><strong>Methods: </strong>We implemented a web-based, self-hosted platform for managing educational courses and developing AI-driven bioinformatics tools. This platform is based on the Git ecosystem as a decentralized version control system which facilitates collaboration among distributed teams.</p><p><strong>Results: </strong>The implemented platform enables comprehensive repository management and secure user access control, providing an integrated collaborative environment. It also supports educational activities, including course definition, sharing of teaching materials, and assignment submission. Within this framework, students and instructors can effectively coordinate tasks and store detailed project documentation throughout all stages of tool development. As a practical example, a new bioinformatics software package was successfully developed using this platform, demonstrating the system's effectiveness in supporting both educational and research-oriented software development.</p><p><strong>Conclusion: </strong>Our platform provides a secure and scalable environment for teaching and research in bioinformatics and AI in medicine. It provides advanced collaboration features that support learning activities and the development of computational tools, essential for improving both educational and research outcomes. We recommend its use in academic and research settings to strengthen the required skills and practices for advances in medicine and biology.</p>","PeriodicalId":37680,"journal":{"name":"Journal of Medical Signals & Sensors","volume":"16 ","pages":"21"},"PeriodicalIF":1.8,"publicationDate":"2026-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13557523/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148921245","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Presentation of a New Averaging Method for Improvement of Systematic and Random Errors in Radiotherapy of Chest and Abdomen Cancer Using Electronic Portal Imaging Device.","authors":"Mohammad Reza Asgari, Mahdi Ghorbani, Mehdi Khosravi, Ahmad Shakeri, Mohamad Pursamimi, Meysam Tavakoli","doi":"10.4103/jmss.jmss_35_25","DOIUrl":"10.4103/jmss.jmss_35_25","url":null,"abstract":"<p><strong>Background: </strong>This study aims to introduce and evaluate a new averaging method to decrease systematic and random errors using an electronic portal imaging device (EPID) in radiotherapy of chest and abdomen cancers.</p><p><strong>Methods: </strong>Thirty-three patients with tumors in different chest and abdominal organs were evaluated. A novel averaging method for decreasing systematic and random errors was introduced and compared with conventional methods in radiotherapy of chest and abdomen cancers. In method 1, averaging by EPID was not performed. While in method 2, EPID sessions that had outlier data were also included in the averaging. The new method (method 3) included, removing the EPID sessions that had high random errors, increasing the number of EPID sessions, and averaging from more accurate data. Pair-sample <i>t</i>- and Chi-square tests were used for comparisons of results from the three methods.</p><p><strong>Results: </strong>For comparison between optimized averaging (method 3) and 3-first fraction averaging (method 2) in terms of individual mean set-up error, the <i>P</i> = 0.00, which indicates a significant difference between these two methods. In the method 3, the three-dimensional displacement vector individual mean set-up errors were lower than the other methods. Moreover, in this method, only one patient has a systematic error greater than the gross error, whereas that of methods 2 and 3 accounted for 4 and 10 patients, respectively.</p><p><strong>Conclusion: </strong>The optimized averaging method introduces a great improvement in chest and abdomen cancers radiotherapy by decreasing systematic errors and increasing treatment accuracy.</p>","PeriodicalId":37680,"journal":{"name":"Journal of Medical Signals & Sensors","volume":"16 ","pages":"19"},"PeriodicalIF":1.8,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13354117/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148424657","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fractional Active Contour Model with Adaptive Hessian Weighting for Mammogram Segmentation.","authors":"Ruhollah Motamedi, Nasser Aghazadeh, Mahdi Hashemzadeh, Parisa Noras","doi":"10.4103/jmss.jmss_108_25","DOIUrl":"10.4103/jmss.jmss_108_25","url":null,"abstract":"<p><strong>Background: </strong>Accurate segmentation of breast masses in mammographic images is a crucial step in computer-aided diagnosis (CAD) systems. However, mammograms are often affected by noise, low contrast, weak boundaries, and intensity inhomogeneity, which significantly challenge reliable segmentation.</p><p><strong>Methods: </strong>In this study, a novel variational active contour model is proposed by integrating spatially adaptive fractional-order enhancement with Hessian-based curvature weighting. A spatially varying fractional-order map is first derived from the image gradient to selectively enhance texture details and weak edges. Based on the enhanced image, second-order structural information is extracted through the Hessian matrix, whose eigenvalues are employed to construct an adaptive edge-stopping function sensitive to local curvature. To further address intensity inhomogeneity, a local region-fitting energy term inspired by the region-scalable fitting model is incorporated into the formulation. The proposed model is implemented within a level-set framework, ensuring numerical stability and topological flexibility.</p><p><strong>Results: </strong>The proposed method was evaluated on two publicly available mammogram datasets, namely INbreast and CBIS-DDSM. Quantitative comparisons using the Dice Similarity Coefficient (DSC) and Hausdorff Distance demonstrate that the proposed approach consistently outperforms classical active contour models and achieves competitive performance compared to learning-based methods, particularly in challenging cases involving low contrast and irregular lesion boundaries.</p><p><strong>Conclusions: </strong>The proposed fractional active contour model provides a robust, fully unsupervised, and interpretable solution for mammographic mass segmentation. By jointly exploiting adaptive fractional enhancement, Hessian-based curvature information, and local intensity fitting, the method effectively addresses key challenges inherent to mammogram images, making it a promising tool for assisting breast cancer detection in CAD systems.</p>","PeriodicalId":37680,"journal":{"name":"Journal of Medical Signals & Sensors","volume":"16 ","pages":"17"},"PeriodicalIF":1.8,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13354142/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148424469","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Analysis of Event-related Potential Dynamics Reveals Lavender Aroma Modulates Cognitive Flexibility: An Electroencephalography Investigation Based on Task Switching Paradigms.","authors":"Reyhaneh Afghan, Hamid Soltani Zangbar, Soomaayeh Heysieattalab, Abbas Ebrahimi-Kalan, Amin Besharat, Nasser Samadzadehaghdam","doi":"10.4103/jmss.jmss_100_25","DOIUrl":"10.4103/jmss.jmss_100_25","url":null,"abstract":"<p><strong>Background: </strong>The therapeutic potential of lavender aroma is well documented, with proven results in reducing stress and facilitating mental repose. However, its effects on specific domains of executive function, such as cognitive flexibility, have not been sufficiently studied. This gap is significant for understanding how nonpharmacological interventions can influence complex cognition.</p><p><strong>Methods: </strong>This study investigates these effects by employing task-switching paradigms to assess cognitive flexibility. The neural correlates were measured via electroencephalography (EEG), focusing on P300 event-related potentials (ERPs). P300 latency and amplitude, a key ERP component reflecting cognitive processing, were measured at EEG electrodes Cz and FCz.</p><p><strong>Results: </strong>The data revealed that lavender inhalation induced specific neural changes. Following lavender inhalation, the P300 peak amplitude at the Cz electrode increased during switch trials, which could indicate improved cognitive control in demanding situations. At the FCz electrode, the latency of P300 was significantly reduced during switch trials, which could suggest the importance of this area in processing speed.</p><p><strong>Conclusion: </strong>The findings of the current study verify the neuromodulatory effects of lavender aroma in improving processes related to cognitive flexibility. This provides a physiological basis for its benefits, suggesting it could be effectively applied in future clinical interventions for cognitive enhancement.</p>","PeriodicalId":37680,"journal":{"name":"Journal of Medical Signals & Sensors","volume":"16 ","pages":"18"},"PeriodicalIF":1.8,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13354138/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148424479","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Gluteus Maximus-gastrocnemius versus Quadriceps Biofeedback Neuromuscular Training during Gait in Knee Osteoarthritis: Development and Feasibility of a Novel Concept with a Pilot Study.","authors":"Khosro Khademi Kalantari, Soheila Ganjeh, Khorshid Bijari, Mohammad Mohsen Roostayi, Hamidreza Aslani, Alireza Akbarzadeh Baghban, Zahra Sadat Rezaeian","doi":"10.4103/jmss.jmss_140_25","DOIUrl":"10.4103/jmss.jmss_140_25","url":null,"abstract":"<p><strong>Background: </strong>The neuromuscular training is a subgroup of functional exercises. The effectiveness of the neuromuscular training substantially improves in combination with electromyography (EMG) biofeedback. The present study aimed to investigate the effect of a new model of neuromuscular training in knee osteoarthritis (KOA).</p><p><strong>Methods: </strong>This pilot, parallel randomized-clinical-trial involved 10 participants with moderate KOA, who were randomly assigned into either neuromuscular training for the gluteus maximus and gastrocnemius (pro-group = 5) or the quadriceps (against-group = 5). Muscle activity in the feedback phase of gait (EMG), pain (Visual Analog Scale [VAS] and knee injury and osteoarthritis outcome score [KOOS]-pain score), and function (average walking speed and KOOS score) were assessed at baseline, immediately after the first treatment session (except KOOS), after 10 sessions of intervention, and after a 3-month of follow-up.</p><p><strong>Results: </strong>After the first treatment session, pain slightly (5.71%) increased in the against-group, whereas decreased by 6.45% in the Progroup; after the 10th session and 3 months, all variables in both groups improved, with a slight extra positive difference in the Pro-group. However, after 3 months, the percentage of changes in the Progroup was greater than that of the against-group. The pain intensity based on VAS (-96.8%), pain (+101.30%), and quality of life (+145.41%) scores of the KOOS questionnaire showed nearly 100% or even greater improvement in the pro-group.</p><p><strong>Conclusion: </strong>Retraining quadriceps and gastrocnemius and gluteus maximus using biofeedback during gait seems promising in KOA, although the Progroup apparently experienced considerably greater clinical improvement. The study provides preliminary evidence of the clinical feasibility of a novel neuromuscular training paradigm for KOA based on biofeedback during gait.</p><p><strong>Trial registration: </strong>This study was registered under the International Randomized Controlled Trial Number registry on November 20, 2023.</p>","PeriodicalId":37680,"journal":{"name":"Journal of Medical Signals & Sensors","volume":"16 ","pages":"15"},"PeriodicalIF":1.8,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13354130/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148424681","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Creation of a Standardized Myocardial Perfusion Single-photon Emission Computed Tomography Database to Enhance Coronary Artery Disease Research.","authors":"Abdurrahim Rahimian, Masoud Moslehi, Hossein Rabbani, Mahnaz Etehadtavakol, Tohid Dehghani","doi":"10.4103/jmss.jmss_104_25","DOIUrl":"10.4103/jmss.jmss_104_25","url":null,"abstract":"<p><strong>Background: </strong>Coronary artery disease (CAD) remains the leading global cause of cardiovascular mortality. Although single-photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI) is widely used, progress in artificial intelligence (AI)-based diagnostic tools is constrained by the limited availability of modern, high-quality, and consistently labeled imaging datasets.</p><p><strong>Methods: </strong>We retrospectively analyzed 144 rest-stress MIBI (99m Tc-methoxy isobutyl isonitrile) SPECT MPI studies acquired using ASNC/EANM-compliant protocols. Images were reconstructed using filtered back projection and independently interpreted by three nuclear cardiology specialists. Clinical, demographic, and imaging variables were analyzed using SPSS v26, with <i>P</i> < 0.05 defining statistical significance.</p><p><strong>Results: </strong>Significant sex-specific differences in CAD presentation were observed. Women with CAD were older and more frequently demonstrated perfusion patterns compatible with microvascular dysfunction, whereas men exhibited larger territorial defects. Diabetes prevalence was significantly higher among CAD-positive patients, whereas smoking patterns differed markedly by sex. Family history of CAD was significantly more common among CAD-positive subjects. Perfusion abnormalities correlated strongly with cumulative cardiovascular risk burden. Interobserver agreement was excellent (Intraclass correlation coefficient (ICC) = 0.87).</p><p><strong>Conclusions: </strong>We introduce a rigorously standardized, clinically annotated SPECT MPI dataset tailored for developing and validating explainable AI models in nuclear cardiology. Its sex-stratified structure supports research into CAD phenotypes and enhances reproducibility.</p><p><strong>Public access: </strong>https://misp.mui.ac.ir/fa/t2-dual-head-spectct-coronary-artery-disease-cad.</p>","PeriodicalId":37680,"journal":{"name":"Journal of Medical Signals & Sensors","volume":"16 ","pages":"16"},"PeriodicalIF":1.8,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13354128/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148424472","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Region-specific Transcriptomic Signatures in Alzheimer's Disease: A Meta-analysis of Vulnerable Brain Regions Reveals MicroRNA-hub Gene Regulatory Networks.","authors":"Khojaste Rahimi Jaberi, Shayan Khalili Alashti, Sedighe Hooshmandi, Pooya Vatankhah, Maryam Rohani Haghighi, Amir Savardashtaki, Hadi Aligholi, Abbas Rahimi Jaberi","doi":"10.4103/jmss.jmss_70_25","DOIUrl":"10.4103/jmss.jmss_70_25","url":null,"abstract":"<p><strong>Background: </strong>Alzheimer's disease (AD) is characterized by progressive neurodegeneration in regionally vulnerable brain areas, yet molecular insights into early pathogenic mechanisms remain limited.</p><p><strong>Methods: </strong>We conducted a meta-analysis of transcriptomic datasets from brain regions affected in early-to-moderate AD - including entorhinal cortex, CA1 hippocampus, angular gyrus, and frontal cortex synaptoneurosomes - using data from seven mRNA and one microRNA (miRNA) microarray studies (GSE16759, GSE110226, GSE37264, GSE26972, GSE36980, GSE37263, GSE39420, and GSE157239). Preprocessing included background correction, log₂ transformation, quantile normalization, and batch correction via ComBat. Differentially expressed features were defined as false discovery rate <0.05 and | logFC| ≥ 1.23 (genes) or ≥ 2 (miRNAs).</p><p><strong>Results: </strong>We identified 172 differentially expressed genes (122 upregulated and 50 downregulated) and 82 significant miRNAs. Hub genes included Inositol-trisphosphate 3-kinase B (<i>ITPKB</i>), Synaptotagmin 1, Dystrobrevin alpha (<i>DTNA</i>), X Inactive Specific Transcript, and Regulator of G protein signaling 4 (<i>RGS4</i>). Functional enrichment highlighted calcium signaling, synaptic failure, and neuroinflammation. Notably, hsa-miR-30d-5p was predicted to target both <i>ITPKB</i> and <i>DTNA</i>, suggesting a regulatory axis linking miRNA dysregulation to calcium dyshomeostasis. Receiver operating characteristic analysis revealed that only <i>RGS4</i> showed moderate discriminative capacity (area under the curve [AUC] =0.70), while other hub genes (e.g., <i>ITPKB</i>, AUC = 0.40) exhibited below-chance performance, underscoring the limitations of single-gene classifiers in postmortem tissue.</p><p><strong>Conclusion: </strong>This study provides mechanistic hypotheses - rather than diagnostic biomarkers - by uncovering region-specific, miRNA-mediated regulatory networks in AD-affected brain tissues. Future validation in accessible biofluids is essential before clinical translation.</p>","PeriodicalId":37680,"journal":{"name":"Journal of Medical Signals & Sensors","volume":"16 ","pages":"14"},"PeriodicalIF":1.8,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13354135/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148424938","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Efficient Techniques Based on Sparse Representation for Classifying High-dimensional Multiclass Microarray Data.","authors":"Maliheh Miri, Mohammad Taghi Sadeghi, Vahid Abootalebi","doi":"10.4103/jmss.jmss_52_25","DOIUrl":"10.4103/jmss.jmss_52_25","url":null,"abstract":"<p><strong>Background: </strong>Sparse representation (SR) has shown strong performance in classification tasks, particularly for high-dimensional data such as microarray gene expression profiles. These datasets present significant challenges due to their high dimensionality and limited sample size, which often hinder the performance of conventional classifiers.</p><p><strong>Methods: </strong>SR addresses this by expressing each signal as a linear combination of a small subset of training samples, reducing computational complexity and improving accuracy. However, using all training samples in the dictionary increases computational cost. This study explores several SR-based classifiers to address microarray data classification, focusing on dictionary construction strategies and sparse coding algorithms.</p><p><strong>Results: </strong>Experimental results on the 14-Tumors dataset show that selecting a subset of representative atoms and applying the SL0 algorithm significantly improves both speed and classification accuracy.</p><p><strong>Conclusions: </strong>These findings highlight the potential of SR approaches for effective and efficient classification of high-dimensional biological data.</p>","PeriodicalId":37680,"journal":{"name":"Journal of Medical Signals & Sensors","volume":"16 ","pages":"13"},"PeriodicalIF":1.8,"publicationDate":"2026-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13274944/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148279083","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Supervised Volumetric Segmentation of White and Gray Matter from Brain Positron Emission Tomography Images Using Magnetic Resonance Labels.","authors":"Kiarash Danesh, Alireza Karimian, Peyman Sharifian, Hossein Arabi","doi":"10.4103/jmss.jmss_49_25","DOIUrl":"10.4103/jmss.jmss_49_25","url":null,"abstract":"<p><strong>Background: </strong>This study evaluates the performance of five deep convolutional neural networks (DCNNs) for supervised segmentation of white matter (WM) and gray matter (GM) in brain positron emission tomography (PET) images using label maps derived from corresponding magnetic resonance (MR) images. The goal is to reduce dependency on hybrid PET/Magnetic resonance imaging (MRI) by extracting anatomical information solely from PET images, thereby providing a potential pathway for MRI-free partial volume correction (PVC).</p><p><strong>Methods: </strong>A total of 300 PET images and their corresponding MR-derived labels from the OASIS-3 dataset were used. The five DCNNs, including the UNet, RegUNet, VNet, SegResNet, and HighResNet, were implemented within the Medical Open Network for Artificial Intelligence (MONAI) framework. The networks were evaluated using Dice score and Intersection over Union (IoU) metrics.</p><p><strong>Results: </strong>Among the networks, VNet demonstrated superior performance for GM and WM segmentation, achieving Dice scores of 61.18% and 76.23%, and IoU scores of 44.15% and 61.62%, respectively. UNet and VNet showed competitive performance for WM segmentation, with no statistically significant differences between them.</p><p><strong>Conclusions: </strong>These findings provide insights into the performance of DCNNs for PET image segmentation, highlighting VNet's capability and emphasizing the potential for optimizing segmentation techniques for PVC applications.</p>","PeriodicalId":37680,"journal":{"name":"Journal of Medical Signals & Sensors","volume":"16 ","pages":"12"},"PeriodicalIF":1.8,"publicationDate":"2026-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13274948/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148279120","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}