Biomedical Signal Processing and Control最新文献

{"title":"Medical image denoising using optimal attention block-based pyramid denoising network","authors":"Vaibhav Jain ,&nbsp;Ashutosh Datar ,&nbsp;Yogendra Kumar Jain","doi":"10.1016/j.bspc.2025.107794","DOIUrl":"10.1016/j.bspc.2025.107794","url":null,"abstract":"<div><div>Over the past two decades, with advancement of computing technologies and applications has significantly enhanced medical imaging and diagnostic processes. The integration of these advanced computing resources, has improved the diagnosis and treatment of various diseases. However, medical imaging technologies often involve noise that makes the diagnosis process more challenging to distinguish among different areas or objects with medical images. Noise and artifacts can easily infiltrate medical images. This results in denoised images to lose some important information. Extracting useful information from noisy images it is a major challenge. To tackle this issue denoising is a crucial pre-processing step to preserve high-quality recovered images. This paper is dedicated to solving such an issue and proposes a hybrid model to handle it. The paper presented an optimal attention block (OAB) based Pyramid denoising Network (OABPDN) whose function is to estimate the optimal co-efficient for image denoising. The model is composed of an attention block that extracts optimal weight for noise component estimation using cuckoo search optimization (CSO). The entire OABPDN is composed of three processing units i.e., optimal pre-processing block (OPB), multi-scale pyramidal network integrated with OAB, and pyramidal feature selection block (PFSB). The result was performed for different noise scales and with different types of noise such as Gaussian, speckle and Poisson noise. These noises are induced in the dataset artificially. The experiment analysis was performed on different types of medical images. The CHASEDB1, MRI, and Lumbar Spine datasets were used for result evaluation. The result was evaluated with varying OAB Parameters, varying noise levels, and noise type. Then the proposed OABPDN was compared to existing models and it was observed that the proposed OABPDN model outperforms better. The model shows approx. 2–3 % improvement of PSNR and SSIM over existing state-of-art models.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"108 ","pages":"Article 107794"},"PeriodicalIF":4.9,"publicationDate":"2025-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143859613","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 multi-layered stacked classifier developed for diagnosis of Parkinson’s disease and SWEDD patients using fusion of multimodal data","authors":"Nikita Aggarwal ,&nbsp;Barjinder Singh Saini ,&nbsp;Savita Gupta","doi":"10.1016/j.bspc.2025.107924","DOIUrl":"10.1016/j.bspc.2025.107924","url":null,"abstract":"<div><div>Early detection of Parkinson’s disease (PD) is difficult due to overlapping with the common symptoms of other neuro-disorders. One of the most prominent of these related diseases is SWEDD (scans without evidence of dopamine deficit), which is considered clinically similar to PD and also has normal dopamine transporter scans. Therefore, there is a pressing need for a reliable method for distinguishing PD from SWEDD and related disorders. To handle this problem, the association between PD and SWEDD has been explored using the fusion of features based on multimodal data (biological, clinical, and imaging). First, the data is normalized by implementing the min–max normalization. Subsequently, feature selection and data-balancing strategies are applied to select optimal features and overcome the data imbalance issue. In addition, a multi-layered stacking (MULS) classifier of three layers is proposed for classification. Also, Bayesian optimization and 5-fold nested stratified cross-validation for hyperparameter tuning are applied on each layer of the MULS classifier. The performance of the developed classifier is estimated using the best feature set against three binary classifications. From the outcomes, it has been observed that the MULS classifier achieved better results for classification between PD and SWEDD compared to the methods in the literature. The results yielded are 97.38% accuracy, 96.21% f1-score, 98.78% sensitivity, 98.47% precision, and 98.21% area under the curve. Furthermore, the impact of multimodal fusion features is analyzed, and also the proposed model is validated with the independent datasets. Hence, the suggested method is believed to help healthcare professionals analyze diseases early.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"108 ","pages":"Article 107924"},"PeriodicalIF":4.9,"publicationDate":"2025-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143859614","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":"Advanced explainable AI-driven biomarker identification for early breast cancer detection using peripheral blood mononuclear cells: Insights into prognostic biomarkers","authors":"Azam jafarabadi ,&nbsp;Elahe Sadat Abdolkarimi","doi":"10.1016/j.bspc.2025.107910","DOIUrl":"10.1016/j.bspc.2025.107910","url":null,"abstract":"<div><div>Breast cancer is one of the leading causes of death worldwide. Despite advances in treatment, its increasing prevalence is a serious concern. Peripheral blood mononuclear cells (PBMCs) undergo gene expression changes when interacting with tumors and can be considered as promising biomarkers for early detection. This study aimed to identify potential biomarkers for breast cancer using explainable artificial intelligence (XAI) and machine learning models. Two datasets, GSE27562 and GSE47862, included healthy individuals and breast cancer patients. After careful preprocessing and data fusion, several machine learning models, including AdaBoost, XGBoost, Random Forest, and Decision Tree, were tested. The AdaBoost model achieved the highest accuracy of 98%. Using SHAP values, ten key genes that had the greatest impact on the model prediction were identified: MRPL3, SLC36A4, COMT, HAAO, KCTD10, FCHO1, RND2, RBM7, LBX1, and LTB4R. Pathway and functional analysis showed that these genes are involved in important processes such as protein metabolism and signal transduction and have high potential as biomarkers. Survival analysis was used to investigate the role of these genes in breast cancer prognosis, and Protein–Protein Interaction (PPI) analysis provided insights into the relationship and gene interaction networks. The findings of this study emphasize the high importance of PBMCs as a non-invasive tool for breast cancer prognosis and indicate that, given the high accuracy, interpretability, and potential of this method in clinical application, it can be used to transform cancer prognosis and develop therapeutic strategies<strong>.</strong></div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"108 ","pages":"Article 107910"},"PeriodicalIF":4.9,"publicationDate":"2025-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143859615","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":"Graph embedding dimensionality reduction combined with improved APO optimized kELM for pneumonia recognition","authors":"Wenhao Lai ,&nbsp;Duoduo Liu ,&nbsp;Jialong Yang ,&nbsp;Weijin Qian ,&nbsp;Lei Guo ,&nbsp;Jiaojiao Wu ,&nbsp;Haifeng Zhou","doi":"10.1016/j.bspc.2025.107909","DOIUrl":"10.1016/j.bspc.2025.107909","url":null,"abstract":"<div><div>Pneumonia is a significant global public health concern, and accurate recognition is essential for improving global health outcomes. In this study, we propose an improved Artificial Protozoa Optimizer Kernel Extreme Learning Machine (iAPO-kELM) combined with Graph Embedding Dimensionality Reduction (GEDR) method for pneumonia detection. Specifically, we reduce the dimension of lung X-ray image data based on Graph Embedding Extreme Learning Machine (ELM). To enhance recognition efficiency and accuracy, the reduced-dimensionality data is prioritized, and the APO algorithm is improved to optimize the parameters of the kELM algorithm. Additionally, we study the kELM classification performance for pneumonia using different kernel functions. To validate the superiority and reliability of the proposed method, we compare it with other optimization algorithms, dimensionality reduction methods, and classification algorithms. Multiple evaluation metrics, including Precision, Recall, and F1 score, are used for assessment. The results of the five-fold cross-validation experiment show that iAPO-kELM combined with GEDR achieves Precision, Recall, and F1 scores of 0.9716, 0.9714, and 0.9715, respectively, outperforming competitive algorithms. These findings suggest that the proposed approach can assist radiologists in diagnosing pneumonia from chest X-ray images effectively.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"108 ","pages":"Article 107909"},"PeriodicalIF":4.9,"publicationDate":"2025-04-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143859611","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":"Evaluating cognitive enhancement through TENS: An EEG classification framework using TQWT, PSR, and ETNet","authors":"Yingfeng Ouyang,&nbsp;Bingo Wing-Kuen Ling","doi":"10.1016/j.bspc.2025.107950","DOIUrl":"10.1016/j.bspc.2025.107950","url":null,"abstract":"<div><div>Transcutaneous Electrical Nerve Stimulation (TENS) is a widely used therapy for cognitive enhancement. Traditionally, its efficacy has been assessed through subjective feedback, which is prone to errors and insufficient for accurately evaluating therapeutic outcomes. To address this limitation, we propose a novel framework for assessing the cognitive enhancement effects of TENS therapy using single channel electroencephalogram (EEG) signals. Specifically, we constructed a new EEG dataset collected during TENS therapy, with TENS serving as the sole experimental variable. EEGs recorded during cognitive tests were classified to assess the therapy’s impact. The classification accuracy reflects the extent to which TENS impact EEGs, providing an objective measure of its effects on cognitive enhancement. The EEGs are first decomposed into multiple components using the Tunable-Q factor Wavelet Transform (TQWT). Subsequently, these components are embedded into a high-dimensional space using Phase Space Reconstruction (PSR). We then design a EEG classification network, ETNet, which efficiently learns both the time–frequency features and nonlinear dynamics extracted in the earlier steps. Numerical simulation results demonstrate the effectiveness of our approach, achieving an average accuracy (ACC) of 91.31%, a standard deviation (STD) of 2.52%, a kappa score of 0.8247, an F1 score of 0.9125, and an AUC of 0.9390. These results surpass those of various existing methods, underscoring the practical utility of our proposed evaluation framework.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"108 ","pages":"Article 107950"},"PeriodicalIF":4.9,"publicationDate":"2025-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143852200","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":"Contact Free human heart rate prediction using LWHRPnet deep learning in real time face and wrist videos","authors":"S. Anusha ,&nbsp;R. Manjith","doi":"10.1016/j.bspc.2025.107930","DOIUrl":"10.1016/j.bspc.2025.107930","url":null,"abstract":"<div><div>The SARS-CoV-2 pandemic has highlighted the critical need for remote health monitoring, a method that is expected to remain a key approach for delivering medical care in the future. On the other hand, contactless monitoring of vital signs, such as heart rate (HR), is highly challenging. This is because the amplitude of the physiological signal is quite uncertain and can be readily distorted due to noise. Noise can originate from a variety of sources, including head motions, variations in light, and acquiring appliances. The detection of heart rate without physical touch will become an important necessity in the future to prevent the further spread of the disease. This paper proposes a novel light-weight deep learning architecture of LWHRPnet for the prediction of human heart rate using real-time captured face and wrist videos. For face video-based HR detection, a face detection algorithm is used to segment the forehead region and an overlayed mean frame of the forehead is used in the learning of LWHRPnet regression CNN. For the wrist, we performed the segmentation of the blood vessel region from the whole hand and predicted the HR from the mean frame of the same via the same LWHRPnet. Two parallel processes of HR detection implemented by face and wrist simultaneously and comparing the results of HR prediction in terms of RMSE. This simulation performance is compared with earlier works. We achieved a very low RMSE of 0.4137 and a prediction accuracy of 99.86%.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"108 ","pages":"Article 107930"},"PeriodicalIF":4.9,"publicationDate":"2025-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143852202","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":"On autoencoders for extracting muscle synergies: A study in highly variable upper limb movements","authors":"Manuela Giraud ,&nbsp;Cristina Brambilla ,&nbsp;Eleonora Guanziroli ,&nbsp;Salvatore Facciorusso ,&nbsp;Lorenzo Molinari Tosatti ,&nbsp;Franco Molteni ,&nbsp;Alessandro Brusaferri ,&nbsp;Alessandro Scano","doi":"10.1016/j.bspc.2025.107940","DOIUrl":"10.1016/j.bspc.2025.107940","url":null,"abstract":"<div><div>The muscle synergy method is a well-established computational approach to motor control in neuroscience. Recently, the hypothesis of linearity adopted by conventional algorithms has been questioned since the non-linearities of the musculoskeletal system may not be captured by linear methods. The scope of this work is to shed further light on the capabilities of autoencoders (AEs) for extracting muscle synergies by targeting a variety of movements covering the upper limb workspace. This approach elicits multiple muscular activations, which are essential to exploit the potential of muscle synergies. We developed two configurations of an autoencoder: a single-plane model trained and tested on the same movement plane and a multiple-plane model trained on all planes but tested on each plane. Electromyographic data were collected from 16 muscles of 15 participants performing reaching movements across 9 targets in 5 planes, and results were compared to the non-negative factorization (NMF). Both synergies and temporal coefficients showed high similarity between AE and NMF (&gt;0.78), indicating that the motor modules extracted with the two methods have the same structure and similar temporal recruitment. Both methods showed a comparable reconstruction accuracy of the input signal (RMSE and R²). The performance of AE decreased with multiple plane training with respect to single plane training due to signal variability. Limitations of this study include the lack of ground truth and unexplored AE configurations. To foster future work, we released an open codebase to provide an easy-to-use code for reproducing our study and for testing new features that may improve the application of the AE (<span><span>https://github.com/cbrambilla/MuscleSynergyExtractionBench-main</span><svg><path></path></svg></span>). Future research will focus on the development of non-linear techniques to extract muscle synergy in different datasets (e.g., lower limbs, full-body movements, patient populations), applying different setting parameters, multi-layer architectures, and activation functions, and incorporating task performance within synergy models.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"108 ","pages":"Article 107940"},"PeriodicalIF":4.9,"publicationDate":"2025-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143854459","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 3D end-to-end multi-task learning network for predicting lymph node metastasis at multiple nodal stations in gastric cancer","authors":"Hao Zhu ,&nbsp;Zhi Yang ,&nbsp;Chang Zheng ,&nbsp;Ping Jiang ,&nbsp;Yi Fang ,&nbsp;Yuejie Xu ,&nbsp;Ying Xiang ,&nbsp;En Xu ,&nbsp;Lei Wang ,&nbsp;Shanhua Bao ,&nbsp;Wenxian Guan ,&nbsp;Xiaoping Zou","doi":"10.1016/j.bspc.2025.107802","DOIUrl":"10.1016/j.bspc.2025.107802","url":null,"abstract":"<div><div>Gastric cancer remains a global health concern with high incidence and mortality rates. Accurate preoperative prediction of lymph node (LN) metastasis is crucial for staging, treatment planning, and prognosis. This study introduces a novel 3D end-to-end lymph node metastasis multi-task learning network (LMML-net) designed to predict LN metastasis across multiple nodal stations in gastric cancer. We analyzed a cohort of 293 patients who underwent gastrectomy with LN dissection. Preoperative CT scans, conducted within two weeks before surgery, were utilized. The LMML-net integrates tumor segmentation and LN metastasis prediction, employing a 3D attention-unet for tumor segmentation and a multi-task learning approach to address metastasis at different nodal stations. LMML-net demonstrated robust predictive performance, achieving AUCs of 0.813, 0.820, and 0.805 for total LN metastasis in training, testing, and validating cohorts, respectively. Notably, the model effectively addressed challenges posed by early gastric cancer and exhibited satisfactory results across various nodal stations. Visualization through GradCam highlighted significant contributions of both tumor and connective tissue areas to the predictions, enhancing the model’s interpretability. The LMML-net exhibits strong predictive capabilities for LN metastasis across multiple stations in gastric cancer, including cases of early-stage disease. This innovative approach holds promise for guiding personalized preoperative treatments and surgical planning, potentially improving patient outcomes in gastric cancer management. Code and models will be available at: <span><span>https://github.com/yangzhi028/LMML-net</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"108 ","pages":"Article 107802"},"PeriodicalIF":4.9,"publicationDate":"2025-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143854458","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":"Physiological datasets in stress and anxiety research: A systematic review","authors":"Juan Pablo Cobá Juárez Pegueros ,&nbsp;Jorge Rodríguez-Arce","doi":"10.1016/j.bspc.2025.107928","DOIUrl":"10.1016/j.bspc.2025.107928","url":null,"abstract":"<div><div>Stress and anxiety have seen a marked increase in reported cases. Traditional research relies on self-reported assessments, often limited by subjectivity and the variability of individual perceptions. To overcome these limitations, there is a growing focus on the use of physiological signals to detect stress and anxiety more objectively. In this context, collecting and analyzing physiological data in studies on stress and anxiety facilitates scientific understanding and provides a robust foundation for testing and validating new detection technologies and methodologies. This systematic review examines 58 state-of-the-art datasets, highlighting key characteristics such as the physiological signals captured, their attributes, population diversity, data acquisition devices, and regulatory compliance. It also highlights significant limitations in the existing datasets, addressing a critical gap in the literature on physiological signal processing. By providing this analysis, the review aims to guide future research efforts toward creating more robust and diverse datasets.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"108 ","pages":"Article 107928"},"PeriodicalIF":4.9,"publicationDate":"2025-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143852201","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":"Low-cost self-constructing multi-objective multi-mode parallel vestibular schwannoma recognition method","authors":"Lei Zhang,&nbsp;Yahong Yu,&nbsp;Yun Li,&nbsp;Fangchen Peng,&nbsp;Hongping Wen","doi":"10.1016/j.bspc.2025.107964","DOIUrl":"10.1016/j.bspc.2025.107964","url":null,"abstract":"<div><div>Convolutional neural networks (CNNs) have achieved great success in the fields of radiology and pathology. The automatic recognition of vestibular schwannoma (VS) based on magnetic resonance imaging (MRI) can significantly enhance the speed and accuracy of disease diagnosis, and reduce the threat of the disease to patients’ lives. At present, patients are diagnosed using contrast enhanced T1-weighted mode images from MRI but there is growing interest in high resolution T2-weighted mode images. However, due to the complex relationship between these two modes, applying a CNN using a simple multi-mode fusion strategy makes it difficult to learn complex information between the modes, and the feature information cannot be well matched and fused. In addition, most CNN hyper-parameters require fine tuning by experts in numerous “trial and error” experiments to achieve better results, and it is difficult to balance multiple objectives such as the model accuracy and training time. The cost of optimization is very expensive. Therefore, we propose a high-performance “non-deep” VS recognition model with dual-mode multi-channel feature perception coupled with a surrogate-assisted multi-objective particle swarm optimization algorithm based on a Kullback–Leibler (KL)-Dropout network to balance multiple objectives while reducing model optimization costs and human influence. Our experimental results showed that the proposed algorithm reached the optimal level in the benchmark test problem. By combining the proposed algorithm with the proposed model, the accuracy was better in the comparison and the amount calculated by the model was controllable, which verified the effectiveness and generalizability of the proposed method.</div></div>","PeriodicalId":55362,"journal":{"name":"Biomedical Signal Processing and Control","volume":"108 ","pages":"Article 107964"},"PeriodicalIF":4.9,"publicationDate":"2025-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143848252","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}