{"title":"Unveiling Alzheimer’s disease through brain age estimation using multi-kernel regression network and magnetic resonance imaging","authors":"Raveendra Pilli, Tripti Goel, R. Murugan","doi":"10.1016/j.cmpb.2025.108617","DOIUrl":"10.1016/j.cmpb.2025.108617","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>Structural magnetic resonance imaging (MRI) studies have unveiled age-related anatomical changes across various brain regions. The disparity between actual age and estimated age, known as the Brain-Predicted Age Difference (Brain-PAD), serves as an indicator for predicting neurocognitive ailments or brain abnormalities resulting from diseases. This study aims to develop an accurate brain age prediction model that can assist in identifying potential neurocognitive impairments.</div></div><div><h3>Methods:</h3><div>The present study implemented a brain age prediction model using a ResNet-50 deep network and a multi-kernel extreme learning machine (MKELM) regression network, relying on MRI images. Kernel methods translate input information into higher-dimensional space by introducing nonlinearity and enabling the model to grasp complicated data patterns. A multi-kernel function combines the Gaussian and polynomial kernels and is incorporated into the brain age regression model. The model effectively utilizes the benefits of both kernel functions to estimate the ages accurately. MRI scans are segmented into gray matter (GM) and white matter (WM) maps preprocessed and extracted of significant features using the ResNet-50 deep network. Extracted features of the WM and GM datasets are fed into the MKELM regression model for brain age prediction.</div></div><div><h3>Results:</h3><div>The proposed age estimation framework achieved 3.06 years of mean absolute error (MAE) and 4.12 years of root mean square error (RMSE) on healthy controls (HC) WM scans, and on GM scans, 2.73 years of MAE and 3.65 years of RMSE values. To further validate the importance of Brain-PAD as a biomarker for identifying brain health conditions, an independent testing dataset of mild cognitive impairment (MCI) and Alzheimer’s disease (AD) subjects age is predicted. The Brain-PAD values for AD subjects’ GM images are significantly higher compared to those of HC and MCI subjects, indicating distinct brain health conditions. Furthermore, variations in GM and WM tissue were identified in AD subjects, revealing that the parahippocampus and corpus callosum were notably affected.</div></div><div><h3>Conclusion:</h3><div>Our findings underscore the potential of Brain-PAD as a significant biomarker for assessing brain health, with implications for early detection of neurocognitive diseases. The developed framework effectively estimates brain age using MRI, contributing valuable insights into the relationship between brain structure and cognitive health.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108617"},"PeriodicalIF":4.9,"publicationDate":"2025-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143095288","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Awais Ahmed , Xiaoyang Zeng , Rui Xi , Mengshu Hou , Syed Attique Shah
{"title":"Enhancing multimodal medical image analysis with Slice-Fusion: A novel fusion approach to address modality imbalance","authors":"Awais Ahmed , Xiaoyang Zeng , Rui Xi , Mengshu Hou , Syed Attique Shah","doi":"10.1016/j.cmpb.2025.108615","DOIUrl":"10.1016/j.cmpb.2025.108615","url":null,"abstract":"<div><h3>Background and objective:</h3><div>In recent times, medical imaging analysis (MIA) has seen an increasing interest due to its core application in computer-aided diagnosis systems (CADs). A modality in MIA refers to a specific technology used to produce human body images, such as MRI, CT scans, or X-rays. Each modality presents unique challenges and characteristics, often leading to imbalances within datasets. This significant challenge impedes model training and generalization due to the varying convergence rates of different modalities and the suppression of gradients in less dominant modalities.</div></div><div><h3>Methods:</h3><div>This paper proposes a novel fusion approach, and we named it Slice-Fusion. The proposed approach aims to mitigate the modality imbalance problem by implementing a “Modality-Specific-Balancing-Factor” fusion strategy. Furthermore, it incorporates an auxiliary (uni-modal) task that generates balanced modality pairs based on the image orientations of different modalities. Subsequently, a novel multimodal classification framework is presented to learn from the generated balanced modalities. The effectiveness of the proposed approach is evaluated through comparative assessments on a publicly available BraTS2021 <span><span>dataset</span><svg><path></path></svg></span>. The results demonstrate the efficiency of Slice-Fusion in resolving the modality imbalance problem. By enhancing the representation of balanced features and reducing modality bias, this approach holds promise for advancing visual health informatics and facilitating more accurate and reliable medical image analysis.</div></div><div><h3>Results:</h3><div>In the experiment section, three diverse experiments are conducted such as i) Fusion Loss Metrics Evaluation, ii) Classification, and iii) Visual Health Informatics. Notably, the proposed approach achieved an F1-Score of (100%, 81.25%) on the training and validation sets for the classification generalization task. In addition to the Slice-Fusion’s out-performance, the study also created a new modality-aligned dataset (a highly balanced and informative modality-specific image collection) that aids further research and improves MIA’s robustness. These advancements not only enhance the capability of medical diagnostic tools but also create opportunities for future innovations in the field.</div></div><div><h3>Conclusion:</h3><div>This study contributes to advancing medical image analysis, such as effective modality fusion, image reconstruction, comparison, and glioma classification, facilitating more accurate and reliable results, and holds promise for further advancements in visual health informatics.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108615"},"PeriodicalIF":4.9,"publicationDate":"2025-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143094869","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"M2OCNN: Many-to-One Collaboration Neural Networks for simultaneously multi-modal medical image synthesis and fusion","authors":"Jian Zhang, Xianhua Zeng","doi":"10.1016/j.cmpb.2025.108612","DOIUrl":"10.1016/j.cmpb.2025.108612","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>Acquiring comprehensive information from multi-modal medical images remains a challenge in clinical diagnostics and treatment, due to complex inter-modal dependencies and missing modalities. While cross-modal medical image synthesis (CMIS) and multi-modal medical image fusion (MMIF) address certain issues, existing methods typically treat these as separate tasks, lacking a unified framework that can generate both synthesized and fused images in the presence of missing modalities.</div></div><div><h3>Methods:</h3><div>In this paper, we propose the Many-to-One Collaboration Neural Network (M2OCNN), a unified model designed to simultaneously address CMIS and MMIF. Unlike traditional approaches, M2OCNN treats fusion as a specific form of synthesis and provides a comprehensive solution even when modalities are missing. The network consists of three modules: the Parallel Untangling Hybrid Network, Comprehensive Feature Router, and Series Omni-modal Hybrid Network. Additionally, we introduce a mixed-resolution attention mechanism and two transformer variants, Coarsormer and ReCoarsormer, to suppress high-frequency interference and enhance model performance.</div><div>M2OCNN outperformed state-of-the-art methods on three multi-modal medical imaging datasets, achieving an average PSNR improvement of 2.4 dB in synthesis tasks and producing high-quality fusion images despite missing modalities. The source code is available at <span><span>https://github.com/zjno108/M2OCNN</span><svg><path></path></svg></span>.</div></div><div><h3>Conclusion:</h3><div>M2OCNN offers a novel solution by unifying CMIS and MMIF tasks in a single framework, enabling the generation of both synthesized and fused images from a single modality. This approach sets a new direction for research in multi-modal medical imaging, with implications for improving clinical diagnosis and treatment.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108612"},"PeriodicalIF":4.9,"publicationDate":"2025-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143094871","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Semi-supervised Strong-Teacher Consistency Learning for few-shot cardiac MRI image segmentation","authors":"Yuting Qiu , James Meng , Baihua Li","doi":"10.1016/j.cmpb.2025.108613","DOIUrl":"10.1016/j.cmpb.2025.108613","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>Cardiovascular disease is a leading cause of mortality worldwide. Automated analysis of heart structures in MRI is crucial for effective diagnostics. While supervised learning has advanced the field of medical image segmentation, it however requires extensive labelled data, which is often limited for cardiac MRI.</div></div><div><h3>Methods:</h3><div>Drawing on the principle of consistency learning, we introduce a novel semi-supervised Strong-Teacher Consistency Network for few-shot multi-class cardiac MRI image segmentation, leveraging largely available unlabelled data. This model incorporates a student–teacher architecture. A multi-teacher structure is introduced to learn diverse perspectives and avoid local optimals when dealing with largely varying cardiac structures and anatomical features. It employs a hybrid loss that emphasizes consistency between student and teacher representations, alongside supervised losses (e.g., Dice and Cross-entropy), tailored to the challenge of unlabelled data. Additionally, we introduced feature-space virtual adversarial training to enhance robust feature learning and model stability.</div></div><div><h3>Results:</h3><div>Evaluation and ablation studies on the MM-WHS and ACDC benchmark datasets show that the proposed model outperforms nine state-of-the-art semi-supervised methods, particularly with limited annotated data. It achieves 90.14% accuracy on MM-WHS and 78.45% accuracy on ACDC at labelling rates of 25% and 1%, respectively. It also highlights its unique advantages over fully-supervised and single-teacher approaches.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108613"},"PeriodicalIF":4.9,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143077919","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yingfan Ma , Mingzhi Yuan , Ao Shen , Xiaoyuan Luo , Bohan An , Xinrong Chen , Manning Wang
{"title":"SeLa-MIL: Developing an instance-level classifier via weakly-supervised self-training for whole slide image classification","authors":"Yingfan Ma , Mingzhi Yuan , Ao Shen , Xiaoyuan Luo , Bohan An , Xinrong Chen , Manning Wang","doi":"10.1016/j.cmpb.2025.108614","DOIUrl":"10.1016/j.cmpb.2025.108614","url":null,"abstract":"<div><h3>Background and Objective</h3><div>Pathology image classification is crucial in clinical cancer diagnosis and computer-aided diagnosis. Whole Slide Image (WSI) classification is often framed as a multiple instance learning (MIL) problem due to the high cost of detailed patch-level annotations. Existing MIL methods primarily focus on bag-level classification, often overlooking critical instance-level information, which results in suboptimal outcomes. This paper proposes a novel semi-supervised learning approach, SeLa-MIL, which leverages both labeled and unlabeled instances to improve instance and bag classification, particularly in hard positive instances near the decision boundary.</div></div><div><h3>Methods</h3><div>SeLa-MIL reformulates the traditional MIL problem as a novel semi-supervised instance classification task to effectively utilize both labeled and unlabeled instances. To address the challenge where all labeled instances are negative, we introduce a weakly supervised self-training framework by solving a constrained optimization problem. This method employs global and local constraints on pseudo-labels derived from positive WSI information, enhancing the learning of hard positive instances and ensuring the quality of pseudo-labels. The approach can be integrated into end-to-end training pipelines to maximize the use of available instance-level information.</div></div><div><h3>Results</h3><div>Comprehensive experiments on synthetic datasets, MIL benchmarks, and popular WSI datasets demonstrate that SeLa-MIL consistently outperforms existing methods in both instance and bag-level classification, with substantial improvements in recognizing hard positive instances. Visualization further highlights the method’s effectiveness in pathology regions relevant to cancer diagnosis.</div></div><div><h3>Conclusion</h3><div>SeLa-MIL effectively addresses key challenges in MIL-based WSI classification by reformulating it as a semi-supervised problem, leveraging both weakly supervised learning and pseudo-labeling techniques. This approach improves classification accuracy and generalization across diverse datasets, making it valuable for pathology image analysis.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108614"},"PeriodicalIF":4.9,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143294124","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Hangzhi He , Hui Zhao , Lifang Li , Hong Yang , Jingjing Yan , Yiwei Yuan , Xiangwen Hu , Yanbo Zhang
{"title":"Non-experimental rapid identification of lower respiratory tract infections in patients with chronic obstructive pulmonary disease using multi-label learning","authors":"Hangzhi He , Hui Zhao , Lifang Li , Hong Yang , Jingjing Yan , Yiwei Yuan , Xiangwen Hu , Yanbo Zhang","doi":"10.1016/j.cmpb.2025.108618","DOIUrl":"10.1016/j.cmpb.2025.108618","url":null,"abstract":"<div><h3>Background and Objective</h3><div>Microbiological culture is a standard diagnostic test that takes a long time to identify lower respiratory tract infections (LRTI) in patients with chronic obstructive pulmonary disease (COPD). This study entailed the development of an interactive decision-support system using multi-label machine learning. It is designed to assist clinical medical staff in the rapid and simultaneous diagnosis of various infections in these patients.</div></div><div><h3>Methods</h3><div>Clinical health record data were collected from inpatients with COPD suspected of having a LRTI. Two major categories of multi-label learning frameworks were integrated with various machine learning algorithms to create 23 predictive models to identify four categories of infection: fungal, gram-negative bacterial, gram-positive bacterial, and multidrug-resistant organism infections. The predictive power of the individual models was tested. Subsequently, the model with the highest comprehensive performance was selected and integrated with SHAP technology to construct a decision support system.</div></div><div><h3>Results</h3><div>Three-thousand-eight-hundred-one subjects participated in this study. LP-RF recorded the highest overall performance, with a Hamming loss of 0.158 (95 %<em>CI</em>: 0.157–0.159) and a samples-precision of 0.894 (95 %<em>CI</em>: 0.891–0.896). The developed diagnostic decision support system generates predicted probability output for each infection category in a specific patient and displays the interpreted output results.</div></div><div><h3>Conclusion</h3><div>The developed multi-label decision support system enables effective prediction of four categories of infections in patients with a history of COPD, and has the potential to curb the overuse of antimicrobial drugs. This system is highly explainable and interactive, providing real-time support in the simultaneous diagnosis of multiple infection categories.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108618"},"PeriodicalIF":4.9,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143104378","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mohd Faruq Abdul Latif , Nik Nazri Nik Ghazali , Shaifulazuar Rozali , Irfan Anjum Badruddin , Sarfaraz Kamangar
{"title":"Evaluation of mandibular advancement surgery efficacy in treating obstructive sleep apnea: A study on turbulence kinetic energy","authors":"Mohd Faruq Abdul Latif , Nik Nazri Nik Ghazali , Shaifulazuar Rozali , Irfan Anjum Badruddin , Sarfaraz Kamangar","doi":"10.1016/j.cmpb.2025.108610","DOIUrl":"10.1016/j.cmpb.2025.108610","url":null,"abstract":"<div><h3>Background and Objective</h3><div>Obstructive sleep apnoea (OSA) is a prevalent sleep disease characterised by recurrent airway obstruction during sleep, resulting in diminished oxygen intake and disrupted sleep patterns. This study investigates the effectiveness of mandibular advancement surgery as a surgical intervention for obstructive sleep apnoea by analysing the postoperative alterations in turbulence kinetic energy (TKE).</div></div><div><h3>Methodology</h3><div>The research involved five subjects receiving mandibular advancement surgery (MAS). The quantification of TKE was performed both before and throughout the method using a combination of computational fluid dynamics (CFD) models and empirical measurements. A suitable grid size of 2.6 million cells for CFD simulations was determined by grid sensitivity analysis and corroborated with physical measurements.</div></div><div><h3>Results</h3><div>The findings indicated a significant increase in TKE for each individual post-procedure, with increments varying from 23 % to 460 %. The elevated TKE indicates a more rapid airflow in the upper airway post-surgery. This is probably attributable to alterations in the airway's morphology resulting from the surgery. The observed rise in speed and turbulence is theoretically supported by Bernoulli's principle, which elucidates the relationship between air flow velocity and the pressure it generates.</div></div><div><h3>Conclusions</h3><div>This study demonstrates that mandibular advancement surgery efficiently alleviates OSA by markedly enhancing airflow and diminishing turbulence in the upper airway post-treatment. The use of physical validation and grid sensitivity analysis in computational fluid dynamics simulations underscores the meticulous technique utilised, offering a comprehensive assessment of the efficacy of the surgical interventions for OSA.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108610"},"PeriodicalIF":4.9,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143294125","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Semantic-driven synthesis of histological images with controllable cellular distributions","authors":"Alen Shahini , Alessandro Gambella , Filippo Molinari , Massimo Salvi","doi":"10.1016/j.cmpb.2025.108621","DOIUrl":"10.1016/j.cmpb.2025.108621","url":null,"abstract":"<div><div>Digital pathology relies heavily on large, well-annotated datasets for training computational methods, but generating such datasets remains challenging due to the expertise required and inter-operator variability. We present SENSE (SEmantic Nuclear Synthesis Emulator), a novel framework for synthesizing realistic histological images with precise control over cellular distributions. Our approach introduces three key innovations: (1) A statistical modeling system that captures class-specific nuclear characteristics from expert annotations, enabling generation of diverse yet biologically plausible semantic content; (2) A hybrid ViT-Pix2Pix GAN architecture that effectively translates semantic maps into high-fidelity histological images; and (3) A modular design allowing independent control of cellular properties including type, count, and spatial distribution. Evaluation on the MoNuSAC dataset demonstrates that SENSE generates images matching the quality of real samples (MANIQA: 0.52 ± 0.03 vs 0.52 ± 0.04) while maintaining expert-verified biological plausibility. In segmentation tasks, augmenting training data with SENSE-generated images improved overall performance (DSC from 79.71 to 84.86) and dramatically enhanced detection of rare cell types, with neutrophil segmentation accuracy increasing from 40.18 to 78.71 DSC. This framework enables targeted dataset enhancement for computational pathology applications while offering new possibilities for educational and training scenarios requiring controlled tissue presentations.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108621"},"PeriodicalIF":4.9,"publicationDate":"2025-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143074242","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Convolutional neural network-based method for the real-time detection of reflex syncope during head-up tilt test","authors":"Minho Choi , Da Young Kim , Ji Man Hong","doi":"10.1016/j.cmpb.2025.108622","DOIUrl":"10.1016/j.cmpb.2025.108622","url":null,"abstract":"<div><h3>Background and objectives</h3><div>Reflex syncope (RS) is the most common type of syncope caused by dysregulation of the autonomic nervous system. Diagnosing RS typically involves the head-up tilt test (HUTT), which tracks physiological signals such as blood pressure and electrocardiograms during postural changes. However, the HUTT is time-consuming and may trigger RS symptoms in patients. Therefore, a real-time monitoring system for RS risk assessment is necessary to enhance medical efficiency and patient convenience. Although several methods have been developed, most depend on manually extracted features from physiological signals, making them susceptible to feature extraction methods and signal noise.</div></div><div><h3>Methods</h3><div>This study introduces a deep learning-based method for real-time RS detection. This method removes the need for manually extracted features by employing an end-to-end architecture consisting of residual and squeeze-and-excitation blocks. The likelihood of RS occurrence was quantified using the proposed method by analyzing a raw blood pressure signal.</div></div><div><h3>Results</h3><div>Data from 1348 patients (1291 normal and 57 with RS) were used to develop and evaluate the proposed method. The area under the receiver operating characteristic curve was 0.972 for RS detection using ten-fold cross-validation. A threshold between zero and one can adjust the performance characteristics of the proposed method. At a threshold of 0.75, the method achieved sensitivity and specificity values of 94.74 and 94.27 %, respectively. Notably, the technique detected RS 165.35 s before its occurrence, on average.</div></div><div><h3>Conclusions</h3><div>The proposed method outperformed conventional methods in RS detection. In addition to its excellent detection performance, this method only requires blood pressure monitoring, reducing reliance on the number of input signals and enhancing its applicability compared to procedures that require multiple signals. These advantages contribute to the development of safer, more convenient, and more efficient RS detection systems.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"264 ","pages":"Article 108622"},"PeriodicalIF":4.9,"publicationDate":"2025-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143578523","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jiahui Zhong , Wenhong Tian , Yuanlun Xie , Zhijia Liu , Jie Ou , Taoran Tian , Lei Zhang
{"title":"PMFSNet: Polarized multi-scale feature self-attention network for lightweight medical image segmentation","authors":"Jiahui Zhong , Wenhong Tian , Yuanlun Xie , Zhijia Liu , Jie Ou , Taoran Tian , Lei Zhang","doi":"10.1016/j.cmpb.2025.108611","DOIUrl":"10.1016/j.cmpb.2025.108611","url":null,"abstract":"<div><h3>Background and objectives:</h3><div>Current state-of-the-art medical image segmentation methods prioritize precision but often at the expense of increased computational demands and larger model sizes. Applying these large-scale models to the relatively limited scale of medical image datasets tends to induce redundant computation, complicating the process without the necessary benefits. These approaches increase complexity and pose challenges for integrating and deploying lightweight models on edge devices. For instance, recent transformer-based models have excelled in 2D and 3D medical image segmentation due to their extensive receptive fields and high parameter count. However, their effectiveness comes with the risk of overfitting when applied to small datasets. It often neglects the vital inductive biases of Convolutional Neural Networks (CNNs), essential for local feature representation.</div></div><div><h3>Methods:</h3><div>In this work, we propose PMFSNet, a novel medical imaging segmentation model that effectively balances global and local feature processing while avoiding the computational redundancy typical of larger models. PMFSNet streamlines the UNet-based hierarchical structure and simplifies the self-attention mechanism’s computational complexity, making it suitable for lightweight applications. It incorporates a plug-and-play PMFS block, a multi-scale feature enhancement module based on attention mechanisms, to capture long-term dependencies.</div></div><div><h3>Results:</h3><div>The extensive comprehensive results demonstrate that our method achieves superior performance in various segmentation tasks on different data scales even with fewer than a million parameters. Results reveal that our PMFSNet achieves IoU of 84.68%, 82.02%, 78.82%, and 76.48% on public datasets of 3D CBCT Tooth, ovarian tumors ultrasound (MMOTU), skin lesions dermoscopy (ISIC 2018), and gastrointestinal polyp (Kvasir SEG), and yields DSC of 78.29%, 77.45%, and 78.04% on three retinal vessel segmentation datasets, DRIVE, STARE, and CHASE-DB1, respectively.</div></div><div><h3>Conclusion:</h3><div>Our proposed model exhibits competitive performance across various datasets, accomplishing this with significantly fewer model parameters and inference time, demonstrating its value in model integration and deployment. It strikes an optimal compromise between efficiency and performance and can be a highly efficient solution for medical image analysis in resource-constrained clinical environments. The source code is available at <span><span>https://github.com/yykzjh/PMFSNet</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108611"},"PeriodicalIF":4.9,"publicationDate":"2025-01-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143074233","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}