Best Pract Res Clin Rheumatol. 2025 Apr 13:102064. doi: 10.1016/j.berh.2025.102064. Online ahead of print.

ABSTRACT

This update addresses new developments in imaging of axial spondyloarthritis from the past 5 years. These have focused mostly on enhanced CT and MRI-based technologies that bring greater precision to the assessment of both inflammatory and structural lesions in the sacroiliac joint. An international consensus has recommended a 4-sequence MRI for routine diagnostic evaluation of the sacroiliac joint aimed at depicting the location and extent of inflammation as well as an erosion-sensitive sequence for structural damage. The latter include high resolution thin slice sequences that accentuate the interface between subchondral bone and the overlying cartilage and joint space as well as synthetic CT, a deep learning-based technique that transforms certain MRI sequences into images resembling CT. Algorithms based on deep learning derived from plain radiographic, CT, and MRI datasets are increasingly more accurate at identifying sacroiliitis and individual lesions observed on images of the sacroiliac joints and spine.

PMID:40229184 | DOI:10.1016/j.berh.2025.102064

Zhongguo Zhen Jiu. 2025 Apr 12;45(4):413-420. doi: 10.13703/j.0255-2930.20240611-0001. Epub 2025 Jan 7.

ABSTRACT

OBJECTIVE: To develop an artificial intelligence-assisted system for the automatic detection of the features of common 21 auricular points based on the YOLOv8 neural network.

METHODS: A total of 660 human auricular images from three research centers were collected from June 2019 to February 2024. The rectangle boxes and features of images were annotated using the LabelMe5.3.1 tool and converted them into a format compatible with the YOLO model. Using these data, transfer learning and fine-tuning training were conducted on different scales of pretrained YOLO neural network models. The model's performance was evaluated on validation and test sets, including the mean average precision (mAP) at various thresholds, recall rate (recall), frames per second (FPS) and confusion matrices. Finally, the model was deployed on a local computer, and the real-time detection of human auricular images was conducted using a camera.

RESULTS: Five different versions of the YOLOv8 key-point detection model were developed, including YOLOv8n, YOLOv8s, YOLOv8m, YOLOv8l, and YOLOv8x. On the validation set, YOLOv8n showed the best performance in terms of speed (225.736 frames per second) and precision (0.998). On the external test set, YOLOv8n achieved the accuracy of 0.991, the sensitivity of 1.0, and the F1 score of 0.995. The localization performance of auricular point features showed the average accuracy of 0.990, the precision of 0.995, and the recall of 0.997 under 50% intersection ration (mAP50).

CONCLUSION: The key-point detection model of 21 common auricular points based on YOLOv8n exhibits the excellent predictive performance, which is capable of rapidly and automatically locating and classifying auricular points.

PMID:40229149 | DOI:10.13703/j.0255-2930.20240611-0001

NMR Biomed. 2025 May;38(5):e70029. doi: 10.1002/nbm.70029.

ABSTRACT

Previous work indicates evidence that cerebrospinal fluid (CSF) plays a crucial role in brain waste clearance processes and that altered flow patterns are associated with various diseases of the central nervous system. In this study, we investigate the potential of deep learning to predict the distribution in human brain of a gadolinium-based CSF contrast agent (tracer) administered intrathecal. For this, T1-weighted magnetic resonance imaging (MRI) scans taken at multiple time points before and after injection were utilized. We propose a U-net-based supervised learning model to predict pixel-wise signal increase at its peak after 24 h. Performance is evaluated based on different tracer distribution stages provided during training, including predictions from baseline scans taken before injection. Our findings show that training with imaging data from only the first 2-h postinjection yields tracer flow predictions comparable to models trained with additional later-stage scans. Validation against ventricular reflux gradings from neuroradiologists confirmed alignment with expert evaluations. These results demonstrate that deep learning-based methods for CSF flow prediction deserve more attention, as minimizing MR imaging without compromising clinical analysis could enhance efficiency, improve patient well-being and lower healthcare costs.

PMID:40229147 | DOI:10.1002/nbm.70029

Med Phys. 2025 Apr 14. doi: 10.1002/mp.17818. Online ahead of print.

ABSTRACT

BACKGROUND: Accurate and reliable segmentation of esophageal gross tumor volume (GTV) in computed tomography (CT) is beneficial for diagnosing and treating. However, this remains a challenging task because the esophagus has a variable shape and extensive vertical range, resulting in tumors potentially appearing at any position within it.

PURPOSE: This study introduces a novel CNN-transformer-based U-shape model (LRRM-U-TransNet) designed to enhance the segmentation accuracy of esophageal GTV. By leveraging advanced deep learning techniques, we aim to address the challenges posed by the variable shape and extensive range of the esophagus, ultimately improving diagnostic and treatment outcomes.

METHODS: Specifically, we propose a long-range relay mechanism to converge all layer feature information by progressively passing adjacent layer feature maps in the pixel and semantic pathways. Moreover, we propose two ready-to-use blocks to implement this mechanism concretely. The Dual FastViT block interacts with feature maps from two paths to enhance feature representation capabilities. The Dual AxialViT block acts as a secondary auxiliary bottleneck to acquire global information for more precise feature map reconstruction.

RESULTS: We build a new esophageal tumor dataset with 1665 real-world patient CT samples annotated by five expert radiologists and employ multiple evaluation metrics to validate our model. Results of a five-fold cross-validation on this dataset show that LRRM-U-TransNet achieves a Dice coefficient of 0.834, a Jaccard coefficient of 0.730, a Precision of 0.840, a HD95 of 3.234 mm, and a Volume Similarity of 0.143.

CONCLUSIONS: We propose a CNN-Transformer hybrid deep learning network to improve the segmentation effect of esophageal tumors. We utilize the local and global information between shallower and deeper layers to prevent early information loss and enhance the cross-layer communication. To validate our model, we collect a dataset composed of 1665 CT images of esophageal tumors from Sichuan Tumor Hospital. The results show that our model outperforms the state-of-the-art models. It is of great significance to improve the accuracy and clinical application of esophageal tumor segmentation.

PMID:40229138 | DOI:10.1002/mp.17818

BMC Pregnancy Childbirth. 2025 Apr 14;25(1):445. doi: 10.1186/s12884-025-07503-5.

ABSTRACT

BACKGROUND: Circulating cell-free RNA (cfRNA) is gaining recognition as an effective biomarker for the early detection of preeclampsia (PE). However, the current methods for selecting disease-specific biomarkers are often inefficient and typically one-dimensional.

PURPOSE: This study introduces a Pyramid Scene Parsing Network (PSPNet) model to predict PE, aiming to improve early risk assessment using cfRNA profiles.

METHODS: The theoretical maximum Preeclamptic Risk Index (PRI) of patients clinically diagnosed with PE is defined as "1", and the control group (NP) is defined as "0", referred to as the clinical PRI. A data preprocessing algorithm was used to screen relevant cfRNA indicators for PE. The cfRNA expression profiles were obtained from the Gene Expression Omnibus (GSE192902), consisting of 180 normal pregnancies (NP) and 69 preeclamptic (PE) samples, collected at two gestational time points: ≤ 12 weeks and 13-20 weeks. Based on the differences in cfRNA expression profiles, the Calculated Ground Truth values of the NP and PE groups in the sequencing data were acquired (Calculated PRI). The differential algorithm was embedded in the PSPNet neural network and the network was then trained using the generated dataset. Subsequently, the real-world sequencing dataset was used to validate and optimize the network, ultimately outputting the PRI values of the healthy control group and the PE group (PSPNet-based PRI). The model's predictive ability for PE was evaluated by comparing the fit between Calculated PRI (Calculated Ground Truth) and PSPNet-based PRI.

RESULTS: The mean absolute error (MAE) between the Calculated Ground Truth the PSPNet-based PRI was 0.0178 for cfRNA data sampled at ≤ 12 gws and 0.0195 for data sampled at 13-20 gws. For cfRNA data sequenced at ≤ 12 gws and 13-20 gws, the corresponding loss values, maximum absolute errors, peak-to-valley error values, mean absolute errors, and average prediction times per sample were 0.0178 (0.0195).

CONCLUSIONS: The present PSPNet model is reliable and fast for cfRNA-based PE prediction and its PRI output allows for continuous PE risk monitoring, introducing an innovative and effective method for early PE prediction. This model enables timely interventions and better management of pregnancy complications, particularly benefiting densely populated developing countries with high PE incidence and limited access to routine prenatal care.

PMID:40229739 | DOI:10.1186/s12884-025-07503-5

BMC Med Imaging. 2025 Apr 14;25(1):120. doi: 10.1186/s12880-025-01661-w.

ABSTRACT

Polyp segmentation is crucial in computer-aided diagnosis but remains challenging due to the complexity of medical images and anatomical variations. Current state-of-the-art methods struggle with accurate polyp segmentation due to the variability in size, shape, and texture. These factors make boundary detection challenging, often resulting in incomplete or inaccurate segmentation. To address these challenges, we propose DCATNet, a novel deep learning architecture specifically designed for polyp segmentation. DCATNet is a U-shaped network that combines ResNetV2-50 as an encoder for capturing local features and a Transformer for modeling long-range dependencies. It integrates three key components: the Geometry Attention Module (GAM), the Contextual Attention Gate (CAG), and the Multi-scale Feature Extraction (MSFE) block. We evaluated DCATNet on five public datasets. On Kvasir-SEG and CVC-ClinicDB, the model achieved mean dice scores of 0.9351 and 0.9444, respectively, outperforming previous state-of-the-art (SOTA) methods. Cross-validation further demonstrated its superior generalization capability. Ablation studies confirmed the effectiveness of each component in DCATNet. Integrating GAM, CAG, and MSFE effectively improves feature representation and fusion, leading to precise and reliable segmentation results. These findings underscore DCATNet's potential for clinical application and can be used for a wide range of medical image segmentation tasks.

PMID:40229681 | DOI:10.1186/s12880-025-01661-w

J Chem Inf Model. 2025 Apr 14. doi: 10.1021/acs.jcim.5c00476. Online ahead of print.

ABSTRACT

Recently, therapeutic peptides have demonstrated great promise for cancer treatment. To explore powerful anticancer peptides, artificial intelligence (AI)-based approaches have been developed to systematically screen potential candidates. However, the lack of efficient featurization of peptides has become a bottleneck for these machine-learning models. In this paper, we propose a topology-enhanced machine learning model (Top-ML) for anticancer peptide prediction. Our Top-ML employs peptide topological features derived from its sequence "connection" information characterized by spectral descriptors. Our Top-ML model, employing an Extra-Trees classifier, has been validated on the AntiCP 2.0 and mACPpred 2.0 benchmark data sets, achieving state-of-the-art performance or results comparable to existing deep learning models, while providing greater interpretability. Our results highlight the potential of leveraging novel topology-based featurization to accelerate the identification of anticancer peptides.

PMID:40229641 | DOI:10.1021/acs.jcim.5c00476

Commun Med (Lond). 2025 Apr 14;5(1):116. doi: 10.1038/s43856-025-00776-z.

ABSTRACT

BACKGROUND: Endotracheal intubation (ETI) is an emergency procedure performed in civilians and combat casualty care settings to establish an airway. It's crucial that healthcare personnel are proficient in these skills, which traditionally have been evaluated through direct feedback from experts. Unfortunately, this method can be inconsistent and subjective, requiring considerable time and resources.

METHODS: This study introduces a system for assessing ETI skills using video analysis. The system employs advanced video processing techniques, including a 2D convolutional autoencoder (AE) based on a self-supervision model, capable of recognizing complex patterns in videos. A 1D convolutional model enhanced with a cross-view attention module then uses AE features to make assessments. Data for the study was gathered in two phases, focusing first on comparisons between experts and novices, and then examining how novices perform under time constraints with outcomes labeled as either successful or unsuccessful. A separate set of data using videos from head-mounted cameras was also analyzed.

RESULTS: The system successfully distinguishes between experts and novices in initial trials and demonstrates high accuracy in further classifications, including under time pressure and using head-mounted camera footage.

CONCLUSIONS: This system's ability to accurately differentiate between experts and novices instills confidence in its effectiveness and potential to improve training and certification processes for healthcare providers.

PMID:40229550 | DOI:10.1038/s43856-025-00776-z

Commun Biol. 2025 Apr 14;8(1):606. doi: 10.1038/s42003-025-08009-8.

ABSTRACT

We had previously reported a convolutional neural network (CNN) based approach, called the holistic kinetic model (HK model 1), for detecting 5-methylcytosine (5mC) by single molecule real-time sequencing (Pacific Biosciences). In this study, we constructed a hybrid model with CNN and transformer layers, named HK model 2. We improve the area under the receiver operating characteristic curve (AUC) for 5mC detection from 0.91 for HK model 1 to 0.99 for HK model 2. We further demonstrate that HK model 2 can detect other types of base modifications, such as 5-hydroxymethylcytosine (5hmC) and N6-methyladenine (6mA). Using HK model 2 to analyze 5mC patterns of cell-free DNA (cfDNA) molecules, we demonstrate the enhanced detection of patients with hepatocellular carcinoma, with an AUC of 0.97. Moreover, HK model 2-based detection of 6mA enables the detection of jagged ends of cfDNA and the delineation of cellular chromatin structures. HK model 2 is thus a versatile tool expanding the applications of single molecule real-time sequencing in liquid biopsies.

PMID:40229481 | DOI:10.1038/s42003-025-08009-8

Sci Rep. 2025 Apr 14;15(1):12801. doi: 10.1038/s41598-025-97719-4.

ABSTRACT

Current artificial intelligence (AI) trends are revolutionizing medical image processing, greatly improving cervical cancer diagnosis. Machine learning (ML) algorithms can discover patterns and anomalies in medical images, whereas deep learning (DL) methods, specifically convolutional neural networks (CNNs), are extremely accurate at identifying malignant lesions. Deep models that have been pre-trained and tailored through transfer learning and fine-tuning become faster and more effective, even when data is scarce. This paper implements a state-of-the-art Hybrid Learning Network that combines the Progressive Resizing approach and Principal Component Analysis (PCA) for enhanced cervical cancer diagnostics of whole slide images (WSI) slides. ResNet-152 and VGG-16, two fine-tuned DL models, are employed together with transfer learning to train on augmented and progressively resized training data with dimensions of 224 × 224, 512 × 512, and 1024 × 1024 pixels for enhanced feature extraction. Principal component analysis (PCA) is subsequently employed to process the combined features extracted from two DL models and reduce the dimensional space of the feature set. Furthermore, two ML methods, Support Vector Machine (SVM) and Random Forest (RF) models, are trained on this reduced feature set, and their predictions are integrated using a majority voting approach for evaluating the final classification results, thereby enhancing overall accuracy and reliability. The accuracy of the suggested framework on SIPaKMeD data is 99.29% for two-class classification and 98.47% for five-class classification. Furthermore, it achieves 100% accuracy for four-class categorization on the LBC dataset.

PMID:40229435 | DOI:10.1038/s41598-025-97719-4

Sci Rep. 2025 Apr 14;15(1):12811. doi: 10.1038/s41598-025-96459-9.

ABSTRACT

Diffusion tensor imaging (DTI) is essential for assessing brain microstructure but requires long acquisition times, limiting clinical use. Recent deep learning (DL) approaches, such as SuperDTI or deepDTI, improve DTI metrics but demand large, high-quality datasets for training. We propose a self-supervised deep learning with fine-tuning (SSDLFT) framework to reduce training data requirements. SSDLFT involves self-supervised pretraining, which denoises data without clean labels, followed by fine-tuning with limited high-quality data. Experiments using Human Connectome Project data show that SSDLFT outperforms traditional methods and other DL approaches in qualitative and quantitative assessments of DWI reconstructions and tensor metrics. SSDLFT's ability to maintain high performance with fewer training subjects and DWIs presents a significant advancement, enhancing DTI's practical applications in clinical and research settings.

PMID:40229411 | DOI:10.1038/s41598-025-96459-9

Nat Biotechnol. 2025 Apr;43(4):491-500. doi: 10.1038/s41587-025-02626-8.

NO ABSTRACT

PMID:40229366 | DOI:10.1038/s41587-025-02626-8

Sci Rep. 2025 Apr 14;15(1):12812. doi: 10.1038/s41598-025-97297-5.

ABSTRACT

Acute Lymphoblastic Leukemia (ALL) is a life-threatening malignancy characterized by its aggressive progression and detrimental effects on the hematopoietic system. Early and accurate diagnosis is paramount to optimizing therapeutic interventions and improving clinical outcomes. This study introduces a novel diagnostic framework that synergizes the EfficientNet-B7 architecture with Explainable Artificial Intelligence (XAI) methodologies to address challenges in performance, computational efficiency, and explainability. The proposed model achieves improved diagnostic performance, with accuracies exceeding 96% on the Taleqani Hospital dataset and 95.50% on the C-NMC-19 and Multi-Cancer datasets. Rigorous evaluation across multiple metrics-including Area Under the Curve (AUC), mean Average Precision (mAP), Accuracy, Precision, Recall, and F1-score-demonstrates the model's robustness and establishes its superiority over state-of-the-art architectures namely VGG-19, InceptionResNetV2, ResNet50, DenseNet50 and AlexNet . Furthermore, the framework significantly reduces computational overhead, achieving up to 40% faster inference times, thereby enhancing its clinical applicability. To address the opacity inherent in Deep learning (DL) models, the framework integrates advanced XAI techniques, including Gradient-weighted Class Activation Mapping (Grad-CAM), Class Activation Mapping (CAM), Local Interpretable Model-Agnostic Explanations (LIME), and Integrated Gradients (IG), providing transparent and explainable insights into model predictions. This fusion of high diagnostic precision, computational efficiency, and explainability positions the proposed framework as a transformative tool for ALL diagnosis, bridging the gap between cutting-edge AI technologies and practical clinical deployment.

PMID:40229347 | DOI:10.1038/s41598-025-97297-5

Sci Rep. 2025 Apr 14;15(1):12829. doi: 10.1038/s41598-025-96587-2.

ABSTRACT

In light of the challenges currently facing the inheritance of blue calico, including the reduction in the number of inheritors and the contraction of the market, this paper puts forth a stylistic transfer method based on an enhanced cycle consistency generative adversarial network. This approach is designed to facilitate the creation of novel designs for traditional blue calico patterns. To address the shortcomings of existing style transfer models, including the generation of blurry details, poor texture and color effects, and excessive model parameters, we propose the incorporation of the Ghost convolution module and the SRM attention module into the generator network structure. This approach aims to reduce the model parameter quantity and computational cost while enhancing the feature extraction ability of the network. The experimental results demonstrate that the method proposed in this paper not only effectively enhances the content details, texture, and color effects of the generated images, but also successfully combines traditional blue calico with modern daily necessities, thereby enhancing its appeal to young people. This research provides novel insights into the digital protection and innovative development of traditional culture, and illustrates the extensive potential applications of deep learning technology in the field of cultural heritage.

PMID:40229316 | DOI:10.1038/s41598-025-96587-2

PLoS One. 2025 Apr 14;20(4):e0320297. doi: 10.1371/journal.pone.0320297. eCollection 2025.

ABSTRACT

Cardiac auscultation requires the mechanical vibrations occurring on the body's surface, which carries a range of sound frequencies. These sounds are generated by the movement and pulsation of different cardiac structures as they facilitate blood circulation. Subsequently, these sounds are identified as phonocardiogram (PCG). In this research, deep learning models, namely gated recurrent neural Network (GRU), Bidirectional-GRU, and Bi-directional long-term memory (BILSTM) are applied separately to segment four specific regions within the PCG signal, namely S1 (lub sound), the systolic region, S2 (dub sound), and the diastolic region. These models are applied to three well-known datasets: PhysioNet/Computing in Cardiology Challenge 2016, Massachusetts Institute of Technology (MITHSDB), and CirCor DigiScope Phonocardiogram.The PCG signal underwent a series of pre-processing steps, including digital filtering and empirical mode decomposition, after then deep learning algorithms were applied to achieve the highest level of segmentation accuracy. Remarkably, the proposed approach achieved an accuracy of 97.2% for the PhysioNet dataset and 96.98% for the MITHSDB dataset. Notably, this paper represents the first investigation into the segmentation process of the CirCor DigiScop dataset, achieving an accuracy of 92.5%. This study compared the performance of various deep learning models using the aforementioned datasets, demonstrating its efficiency, accuracy, and reliability as a software tool in healthcare settings.

PMID:40228205 | DOI:10.1371/journal.pone.0320297

PLoS One. 2025 Apr 14;20(4):e0320360. doi: 10.1371/journal.pone.0320360. eCollection 2025.

ABSTRACT

Alzheimer's disease (AD) is a neurodegenerative disorder that requires early diagnosis for effective management. However, issues with currently available diagnostic biomarkers preclude early diagnosis, necessitating the development of alternative biomarkers and methods, such as blood-based diagnostics. We propose c-Triadem (constrained triple-input Alzheimer's disease model), a novel deep neural network to identify potential blood-based biomarkers for AD and predict mild cognitive impairment (MCI) and AD with high accuracy. The model utilizes genotyping data, gene expression data, and clinical information to predict the disease status of participants, i.e., cognitively normal (CN), MCI, or AD. The nodes of the neural network represent genes and their related pathways, and the edges represent known relationships among the genes and pathways. Simulated data validation further highlights the robustness of key features identified by SHapley Additive exPlanations (SHAP). We trained the model with blood genotyping data, microarray, and clinical features from the Alzheimer's Neuroimaging Disease Initiative (ADNI). We demonstrate that our model's performance is superior to previous models with an AUC of 97% and accuracy of 89%. We then identified the most influential genes and clinical features for prediction using SHapley Additive exPlanations (SHAP). Our SHAP analysis shows that CASP9, LCK, and SDC3 SNPs and PINK1, ATG5, and ubiquitin (UBB, UBC) expression have a higher impact on model performance. Our model has facilitated the identification of potential blood-based genetic markers of DNA damage response and mitophagy in affected regions of the brain. The model can be used for detection and biomarker identification in other related dementias.

PMID:40228177 | DOI:10.1371/journal.pone.0320360

Environ Health Perspect. 2025 Apr 14. doi: 10.1289/EHP15400. Online ahead of print.

NO ABSTRACT

PMID:40228076 | DOI:10.1289/EHP15400

Magn Reson Med. 2025 Apr 14. doi: 10.1002/mrm.30532. Online ahead of print.

ABSTRACT

PURPOSE: The aim of this study was to develop a saturation transfer MR fingerprinting (ST-MRF) technique using a biophysics model-driven deep learning approach.

METHODS: A deep learning-based quantitative saturation transfer framework was proposed to estimate water, magnetization transfer contrast, and amide proton transfer (APT) parameters plus B0 field inhomogeneity. This framework incorporated a Bloch-McConnell simulator during neural network training and enforced consistency between synthesized MRF signals and experimentally acquired ST-MRF signals. Ground-truth numerical phantoms were used to assess the accuracy of estimated tissue parameters, and in vivo tissue parameters were validated using synthetic MRI analysis.

RESULTS: The proposed ST-MRF reconstruction network achieved a normalized root mean square error (nRMSE) of 9.3% when tested against numerical phantoms with a signal-to-noise ratio of 46 dB, which outperformed conventional Bloch-McConnell fitting (nRMSE of 15.3%) and dictionary-matching approaches (nRMSE of 19.5%). Synthetic MRI analysis indicated excellent similarity (RMSE = 3.2%) between acquired and synthesized ST-MRF images, demonstrating high in vivo reconstruction accuracy. In healthy human brains, the APT pool size ratios for gray and white matter were 0.16 ± 0.02% and 0.13 ± 0.02%, respectively, and the exchange rates for gray and white matter were 101 ± 25 Hz and 131 ± 27 Hz, respectively. The reconstruction network processed the eight tissue parameter maps in approximately 27 s for ST-MRF data sized at 256 × 256 × 9 × 103.

CONCLUSION: This study highlights the feasibility of the deep learning-based ST-MRF imaging for rapid and accurate quantification of free bulk water, magnetization transfer contrast, APT parameters, and B0 field inhomogeneity.

PMID:40228056 | DOI:10.1002/mrm.30532

J Chem Inf Model. 2025 Apr 14. doi: 10.1021/acs.jcim.4c02345. Online ahead of print.

ABSTRACT

The development of new materials is a time-consuming and resource-intensive process. Deep learning has emerged as a promising approach to accelerate this process. However, accurately predicting crystal structures using deep learning remains a significant challenge due to the complex, high-dimensional nature of atomic interactions and the scarcity of comprehensive training data that captures the full diversity of possible crystal configurations. This work developed a neural network model based on a data set comprising thousands of crystallographic information files from existing crystal structure databases. The model incorporates a self-attention mechanism to enhance prediction accuracy by learning and extracting both local and global features of three-dimensional structures, treating the atoms in each crystal as point sets. This approach enables effective semantic segmentation and accurate unit cell prediction. Experimental results demonstrate that for unit cells containing up to 500 atoms, the model achieves a structure prediction accuracy of 89.78%.

PMID:40228012 | DOI:10.1021/acs.jcim.4c02345

J Vis Exp. 2025 Mar 28;(217). doi: 10.3791/67766.

ABSTRACT

Finger gestures are a critical element in human communication, and as such, finger gesture recognition is widely studied as a human-computer interface for state-of-the-art prosthetics and optimized rehabilitation. Surface electromyography (sEMG), in conjunction with deep learning methods, is considered a promising method in this domain. However, current methods often rely on cumbersome recording setups and the identification of static hand positions, limiting their effectiveness in real-world applications. The protocol we report here presents an advanced approach combining a wearable surface EMG and finger tracking system to capture comprehensive data during dynamic hand movements. The method records muscle activity from soft printed electrode arrays (16 electrodes) placed on the forearm as subjects perform gestures in different hand positions and during movement. Visual instructions prompt subjects to perform specific gestures while EMG and finger positions are recorded. The integration of synchronized EMG recordings and finger tracking data enables comprehensive analysis of muscle activity patterns and corresponding gestures. The reported approach demonstrates the potential of combining EMG and visual tracking technologies as an important resource for developing intuitive and responsive gesture recognition systems with applications in prosthetics, rehabilitation, and interactive technologies. This protocol aims to guide researchers and practitioners, fostering further innovation and application of gesture recognition in dynamic and real-world scenarios.

PMID:40227996 | DOI:10.3791/67766