Biomedical Signal Processing and Control最新文献

{"title":"ROPRNet: Deep learning-assisted recurrence prediction for retinopathy of prematurity","authors":"Peijie Huang ,&nbsp;Yiying Xie ,&nbsp;Rong Wu ,&nbsp;Qiuxia Lin ,&nbsp;Nian Cai ,&nbsp;Haitao Chen ,&nbsp;Songfu Feng","doi":"10.1016/j.bspc.2024.107135","DOIUrl":"10.1016/j.bspc.2024.107135","url":null,"abstract":"<div><div>Retinopathy of Prematurity (ROP) recurrence is significant for the prognosis of ROP treatment. In this paper, corrected gestational age at treatment is involved as an important risk factor for the assessment of ROP recurrence. To reveal the complementary information from fundus images and risk factors, a dual-modal deep learning framework with two feature extraction streams, termed as ROPRNet, is designed to assist recurrence prediction of ROP after anti-vascular endothelial growth factor (Anti-VEGF) treatment, involving a stacked autoencoder (SAE) stream for risk factors and a cascaded deep network (CDN) stream for fundus images. Here, the specifically-designed CDN stream involves several novel modules to effectively capture subtle structural changes of retina in the fundus images, involving enhancement head (EH), enhanced ConvNeXt (EnConvNeXt) and multi-dimensional multi-scale feature fusion (MMFF). Specifically, EH is designed to suppress the variations of color and contrast in fundus images, which can highlight the informative features in the images. To comprehensively reveal the inherent medical hints submerged in the fundus images, an adaptive triple-branch attention (ATBA) and a special ConvNeXt with a rare-class sample generator (RSG) were designed to compose the EnConvNeXt for effectively extracting features from fundus images. The MMFF is designed for feature aggregation to mitigate redundant features from several fundus images from different shooting angles, involving a designed multi-dimensional and multi-sale attention (MD-MSA). The designed ROPRNet is validated on a real clinical dataset, which indicate that it is superior to several existing ROP diagnostic models, in terms of 0.894 AUC, 0.818 accuracy, 0.828 sensitivity and 0.800 specificity.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"100 ","pages":"Article 107135"},"PeriodicalIF":4.9,"publicationDate":"2024-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142586396","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":"Automated pericardium segmentation and epicardial adipose tissue quantification from computed tomography images","authors":"Ying Wang ,&nbsp;Ankang Wang ,&nbsp;Lu Wang ,&nbsp;Wenjun Tan ,&nbsp;Lisheng Xu ,&nbsp;Jinsong Wang ,&nbsp;Songang Li ,&nbsp;Jinshuai Liu ,&nbsp;Yu Sun ,&nbsp;Benqiang Yang ,&nbsp;Steve Greenwald","doi":"10.1016/j.bspc.2024.107167","DOIUrl":"10.1016/j.bspc.2024.107167","url":null,"abstract":"<div><h3>Background and Objective</h3><div>Epicardial Adipose Tissue (EAT) is regarded as an independent risk factor for cardiovascular disease, and an increase in its volume is closely associated with disorders such as coronary artery atherosclerosis. Traditional manual and semi-automatic methods for EAT segmentation rely on subjective judgment, resulting in uncertainty and unreliability, which limits their application in clinical practice. Therefore, this study aims to develop a fully automatic segmentation and quantification method to improve the accuracy of EAT assessment.</div></div><div><h3>Methods</h3><div>A Boundary-Enhanced Multi-scale U-Net network with a Convolutional Transformer (BMT-UNet) is developed to segment the pericardium. The BMT-UNet comprises Boundary-Enhanced (BE) modules, Multi-Scale (MS) modules, and a Convolutional Transformer (ConvT) module. The MS and BE modules in the encoding part are designed to capture detailed boundary features and accurately delineate the pericardium boundary by combining multi-scale features with morphological operations, leveraging their complementarity. The ConvT module integrates global contextual information, thereby enhancing overall segmentation accuracy and addressing the issue of internal holes in the segmented pericardial images. The volume of EAT is automatically quantified using standard fat thresholds with a range of −190 to −30 HU.</div></div><div><h3>Results</h3><div>For a Coronary Computed Tomography Angiography (CCTA) dataset which contained 50 patients, the Dice coefficient and Hausdorff distance for the proposed method of pericardial and EAT segmentation are 98.3% ± 0.2%, 5.7±0.8 mm, and 93.9% ± 1.7%, 2.1 ± 0.3 mm, respectively. The linear regression coefficient between the EAT volume segmented and the actual volume is 0.982, and the Pearson correlation coefficient is 0.99. Bland-Altman analysis further confirmed the high consistency between the automated and manual methods. These results demonstrate a significant improvement over existing methods, particularly in terms of segmentation precision and reliability, which are critical for clinical application.</div></div><div><h3>Conclusions</h3><div>This work develops an automated method for quantifying EAT in Computed Tomography (CT) images, and the results agreed closely with expert evaluations. Code is available at: <span><span>https://github.com/wy-9903/BMT-UNet</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"100 ","pages":"Article 107167"},"PeriodicalIF":4.9,"publicationDate":"2024-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142586173","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":"A design of computational stochastic framework for the mathematical severe acute respiratory syndrome coronavirus model","authors":"Atifa Asghar ,&nbsp;Mohsan Hassan ,&nbsp;Zulqurnain Sabir ,&nbsp;Shahid Ahmad Bhat ,&nbsp;Sharifah E Alhazmi","doi":"10.1016/j.bspc.2024.107049","DOIUrl":"10.1016/j.bspc.2024.107049","url":null,"abstract":"<div><div>This study presents the comprehensive investigations into the dynamics of a novel coronavirus infection within a population, which accounts for all potential interactions in the disease’s spread. The solutions of the novel nonlinear infectious disease system are performed stochastically by using the Levenberg-Marquardt Backpropagation neural network. This process contains ten neurons and log-sigmoid transfer function in the hidden layers. The training data is taken as 74%, while the testing and authentication statics are used as 14% and 12%. To assess the precision of the designed solver, a comparison based on the obtained and reference results along with the negligible absolute error up to order fourth to seventh decimal places is performed for each case of the model. Stability and sensitivity analyses reveal the robustness of the model across various parameters. For the reliability, consistency, and correctness of the model across various states, and the numerical analysis with graphical form of the statistical indices based on correlation, error histograms, transition of state, and regression analysis is presented.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"100 ","pages":"Article 107049"},"PeriodicalIF":4.9,"publicationDate":"2024-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142586394","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":"Skin cancer classification based on a hybrid deep model and long short-term memory","authors":"Samira Mavaddati","doi":"10.1016/j.bspc.2024.107109","DOIUrl":"10.1016/j.bspc.2024.107109","url":null,"abstract":"<div><div>Skin cancer classification is an important topic in dermatology and oncology because it provides a framework for diagnosing and managing skin cancer, as well as for research and advocacy efforts. Deep learning-based methods have the potential to improve the efficiency and scalability of skin cancer classification by automatically processing large volumes of images without the need for intervention. The proposed method combines the ResNet50 deep model and long short-term memory (LSTM) network to process sequential data and represent the structural content of lesion texture better to overcome the limitations of a deep learning-based classification algorithm. This hybrid deep classifier, named ResNet50-LSTM, takes advantage of the benefits of both deep networks along with a transfer learning technique which allows a new model to start from a pre-trained model and fine-tune it for the specific task. Three scenarios are demonstrated in this paper that consists, the first one, ResNet50, the second one ResNet50 in combination with transfer learning technique (ResNet50-TL), and the third scenario, (ResNet50-LSTM-TL) deep model. Combining ResNet50, LSTM, and transfer learning techniques can improve the performance of skin cancer classification by allowing the model to take advantage of pre-trained features from a large dataset, analyze sequential features in medical images, and fine-tune them for the specific task of skin cancer classification. The performance of these scenarios is compared with the other deep learning models. The results of the conducted study demonstrate that the proposed third scenario is successful in accurately recognizing various skin cancers, with an impressive accuracy rate of over 99.09%. The findings indicate that the proposed algorithm has the potential to significantly enhance skin cancer classification and by improving their accuracy and efficiency.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"100 ","pages":"Article 107109"},"PeriodicalIF":4.9,"publicationDate":"2024-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142578455","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":"The utility of electroencephalographic measures in obsession compulsion disorder","authors":"Alireza Talesh Jafadideh ,&nbsp;Mehdi Ejtehadi ,&nbsp;Asghar Zarei ,&nbsp;Maryam Ansari Esfeh ,&nbsp;Saeid Yazdi-Ravandi ,&nbsp;Nasrin Matinnia ,&nbsp;Farshid Shamsaei ,&nbsp;Mohammad Ahmadpanah ,&nbsp;Ali Ghaleiha ,&nbsp;Asiyeh Rezaei Niyasar ,&nbsp;Reza Rostami ,&nbsp;Reza Khosrowabadi","doi":"10.1016/j.bspc.2024.107113","DOIUrl":"10.1016/j.bspc.2024.107113","url":null,"abstract":"<div><h3>Background</h3><div>Obsessive-compulsive disorder (OCD) is a potentially serious mental disorder that affects 1–2% of the world population. The OCD patients experience uncontrollable, and recurring thoughts (obsessions), and may feel a need to repeat behaviors (compulsions). In EEG studies, many different features have been investigated regarding their OCD diagnosis capability. However, there is no OCD study evaluating different EEG features with the same conditions and data.</div></div><div><h3>Methods</h3><div>To address the problem, we employed six popular resting-state EEG features, including absolute and relative power, Phase locking value (PLV), Weighted phase lag index (WPLI), approximate entropy, and Higuchi’s fractal dimension, to find out which feature can better discriminate OCDs from healthy controls (CON) under the same conditions and data. All the generated features were normalized using mean and standard deviation of values, calculated from 233 Iranian healthy people. After that, the most informative EEG features, discriminating 39 OCD individuals from age, handedness, and gender-matched, 39 CON were selected and entered into the classification process. In addition, an independent EEG dataset including 23 OCDs and 23 CONs was also used to investigate the consistency of the results.</div></div><div><h3>Results</h3><div>As expected, most of the significant differences were observed at the high frequency bands in Beta I-IV, and Gamma bands. The highest classification accuracies were achieved using the support vector machine applied on the PLV features of the main (94.8 %) and independent dataset (100 %)</div></div><div><h3>Conclusions</h3><div>These findings indicate that functional connectivity-based (PLV) features have a good potential to be used as a biomarker of OCD.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"100 ","pages":"Article 107113"},"PeriodicalIF":4.9,"publicationDate":"2024-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142578454","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":"An Optimized Wasserstein Deep Convolutional Generative Adversarial Network approach for the classification of COVID-19 and pneumonia","authors":"A.B. Rajendra ,&nbsp;B.S. Jayasri ,&nbsp;S. Ramya ,&nbsp;Shruthi Jagadish","doi":"10.1016/j.bspc.2024.107100","DOIUrl":"10.1016/j.bspc.2024.107100","url":null,"abstract":"<div><div>In the context of diagnosing lung disorders like bacterial and viral pneumonia and COVID-19, the challenge of sample scarcity often results in imbalanced datasets, making reliable forecasting difficult. To address this, an Optimized Wasserstein Deep Convolutional Generative Adversarial Network Technique was proposed for the Classification of COVID-19 and Pneumonia (CCP WDCGAN-SOA). The proposed approach utilizes CT scan and X-ray images from two datasets: the COVID-19 Posterior-Anterior Chest Radiography Images Curated Dataset and the COVID QU-Ex Dataset. Due to the imbalance in these datasets, a Label Correlation Guided Borderline Oversampling (LCGBO) method was introduced to balance the classes effectively. Following data balancing, the images undergo pre-processing using Multimodal Hierarchical Graph Collaborative Filtering (MHGCF) for resizing. Subsequently, the processed images are fed into a Wasserstein Deep Convolutional Generative Adversarial Network (WDCGAN) optimized with the Seasons Optimization Algorithm (SOA) to enhance classification accuracy for COVID-19 and pneumonia. The implementation in MATLAB demonstrates that the CCP-WDCGAN-SOA technique significantly outperforms existing methods. Specifically, the proposed approach achieves improvements of 21.5 %, 23 %, and 22.5 % in accuracy, 12.3 %, 17.5 %, and 14 % in recall, and 22.3 %, 27.5 %, and 24 % in specificity compared to DC-CXI-CoviXNet, CPD-CXI-CNN, and ADC-CXI-DFFC Net using the COVID-19 Posterior-Anterior Chest Radiography Images Curated Dataset. Additionally, the proposed method shows gains of 21.52%, 27.05%, and 23.24% in accuracy, 23.71%, 26.45%, and 21.74% in recall, and 28.61%, 22.15%, and 26.44% in specificity over ASC-CXI-LRANet, RCP-MIA-CNN, and AQCD-CR-GAN using the COVID-QU-Ex Dataset.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"100 ","pages":"Article 107100"},"PeriodicalIF":4.9,"publicationDate":"2024-11-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142572450","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":"TIGC-Net: Transformer-Improved Graph Convolution Network for spatio-temporal prediction","authors":"Kai Chen ,&nbsp;Zhengyuan Zhou ,&nbsp;Yao Liu ,&nbsp;Tianjiao Ji ,&nbsp;Weiya Sun ,&nbsp;Chunfeng Yang ,&nbsp;Yang Chen ,&nbsp;Xiao Lu","doi":"10.1016/j.bspc.2024.107024","DOIUrl":"10.1016/j.bspc.2024.107024","url":null,"abstract":"<div><div>Modeling spatio-temporal sequences is an important topic yet challenging for existing neural networks. Most of the current spatio-temporal sequence prediction methods usually capture features separately in temporal and spatial dimensions or employ multiple mutually independent local spatio-temporal graphs to represent a spatio-temporal sequence. The first kind of method mentioned above is difficult to mine the complex spatio-temporal correlations, while the other is limited for the accuracy of long-term predictions. To handle these issues, this paper proposes a Transformer-Improved Graph Convolution Network for spatio-temporal prediction. Specifically, the temporal location encoding method is exploited to derive the spatio-temporal characteristics of the sequence utilizing a spatio-temporal feature fusion network. In addition, a spatio-temporal attention network is developed to enhance the spatio-temporal correlation of the sequence, and the dynamic spatial features of sequence are further extracted through the adaptive graph convolution network. A private dataset and a public dataset are employed to demonstrate the performance of the proposed TIGC-Net. The qualitative and quantitative results show that the proposed TIGC-Net can extract dynamic spatio-temporal properties more effectively, enhance the spatio-temporal correlation of sequences and improve the prediction accuracy compared with four state-of-the-art.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"100 ","pages":"Article 107024"},"PeriodicalIF":4.9,"publicationDate":"2024-11-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142572446","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":"Optimising rooftop photovoltaic adoption in urban landscapes: A system dynamics approach for sustainable energy transitions","authors":"P.U. Poornima ,&nbsp;K. Dhineshkumar ,&nbsp;Chunduri Kiran Kumar ,&nbsp;S. Sumana ,&nbsp;M.V. Rama Sundari ,&nbsp;P. Sivaraman ,&nbsp;Mohammed Shuaib ,&nbsp;A. Rajaram","doi":"10.1016/j.bspc.2024.107071","DOIUrl":"10.1016/j.bspc.2024.107071","url":null,"abstract":"<div><div>Rooftop agriculture for food production and photovoltaic (PV) panels for energy generation are two examples of how urban functional design presents a potential alternative to multi-function urban land-use that may give numerous ecosystem services. In order to find the optimal rooftop usage strategy that takes into account many choice criteria and to comprehend how rooftop solutions affect the layout of urban energy infrastructure, we provide a complete system modeling approach that demonstrates multi-objective optimization of energy systems. With a reduced levelized cost of electricity (LCOE), rooftop photovoltaics have gained considerable traction recently owing to technical, economical, and environmental benefits; this research aims to prove their viability. The suggested PV size and cost factor, taking environmental conditions and shading effects into consideration, were determined using two methods: Quantum Particle Swarm Optimization (PSO) with Q-Learning System. Rooftop photovoltaics system sizing, economic feasibility, and energy efficiency are all affected by the results that are compared. University of Engineering &amp; Technology (UET), a public sector institution, has its main campus in Taxila, where this research was conducted. Situated in northern Pakistan, its appropriate position is advantageous for the research. The lifespan, performance ratio (PR), and decrease of the Rooftop Photovoltaics system’s carbon footprint are among the many additional criteria that are examined. Because of this, installing rooftop photovoltaic systems on government buildings is a more sensible and feasible solution.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"100 ","pages":"Article 107071"},"PeriodicalIF":4.9,"publicationDate":"2024-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142572449","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":"IoT enabled smart healthcare system for COVID-19 classification using optimized robust spatiotemporal graph convolutional networks","authors":"A. Velayudham ,&nbsp;R. Karthick ,&nbsp;A. Sivabalan ,&nbsp;V. Sathya","doi":"10.1016/j.bspc.2024.107104","DOIUrl":"10.1016/j.bspc.2024.107104","url":null,"abstract":"<div><div>Healthcare organizations and academics are paying close attention to the development of smart medical sensors, gadgets, cloud computing, and other health-related technology. To actively diagnose and control the spread of COVID-19, an effectual automated system is required. Therefore, this paper proposes an IoT enabled Smart Healthcare System for COVID-19 Classification Using Optimized Robust Spatiotemporal Graph Convolutional Networks (IoT-RSGCN-SGWOA-CD19). Here, the input images are collected through Chest X-Ray dataset. The input images are preprocessed by utilizing Adaptive two-stage unscented Kalman filter (ATSUKF). Next, the pre-processed images are fed into Two-Dimensional Spectral Graph Wavelets (2DSGW) for extracting features. The extracted features are supplied to the feature selection to select the appropriate features using Clouded Leopard Optimization (CLO). Then, Robust Spatiotemporal Graph Convolutional Network (RSGCN) is proposed to classify the disease as pneumonia, normal and COVID-19. The weight parameter of RSGCN is optimally tuned by Sunflower based Grey Wolf Optimization Algorithm(SFGWOA), improving its accuracy in disease screening and infectious disease categorization. The effectiveness of the proposed IoT-RSGCN-SGWOA-CD19 method is implemented in MATLAB and evaluated through performance metrics, likes accuracy, precision, recall, ROC, AUC, loss. The IoT-RSGCN SGWOA-CD19 method attains 23.64 %, 20.98 % and 24.33 % higher accuracy, 13.24 %, 30.43 % and 28.71 % higher precision and 27.79 %, 23.84 % and 26.62 % higher recall when analyzed with the existing models. The experimental results confirm that the IoT-RSGCN-SGWOA-CD19 method offers a significant advancement in automated COVID-19 screening, with superior classification accuracy and reliability. The proposed system can be a valuable tool in pandemic control by providing rapid and accurate diagnoses.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"100 ","pages":"Article 107104"},"PeriodicalIF":4.9,"publicationDate":"2024-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142573129","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":"FV-DDC: A novel finger-vein recognition model with deformation detection and correction","authors":"Hengyi Ren ,&nbsp;Lijuan Sun ,&nbsp;Jinting Ren ,&nbsp;Ying Cao","doi":"10.1016/j.bspc.2024.107098","DOIUrl":"10.1016/j.bspc.2024.107098","url":null,"abstract":"<div><div>Finger vein recognition has gained widespread attention for personal identification due to its robustness and resistance to forgery. While Convolutional Neural Network (CNN)-based finger vein recognition algorithms have shown promising performance, several challenges remain. Firstly, existing methods often fail to effectively handle complex finger deformations, such as bending and rotation, which frequently occur in real-world applications. Secondly, CNN-based approaches typically require large training datasets, yet the available finger vein datasets are limited in size. To address these challenges, this paper presents a novel CNN-based finger vein recognition algorithm, FV-DDC, incorporating a lightweight finger deformation correction module, FVTN. The FVTN module autonomously learns and corrects finger deformations using matrix transformations, offering a new approach to CNN-based deformation correction. The primary advantages of FV-DDC are twofold: automatic finger deformation correction, which simplifies preprocessing, and data augmentation during deformation correction, reducing the dependency on large datasets. Extensive experiments were conducted on three publicly available datasets to validate the effectiveness of the proposed algorithm. The results show that FV-DDC achieves superior recognition performance, particularly in scenarios involving missing data and deformation interference, with recognition accuracies of 99.62% on HKPU, 99.80% on FV-USM, and 98.74% on SDUMLA.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"100 ","pages":"Article 107098"},"PeriodicalIF":4.9,"publicationDate":"2024-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142572443","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}