Medical & Biological Engineering & Computing最新文献

Multiple sclerosis (MS) is a neurodegenerative disease that impacts retinal layer thickness. Thus, several works proposed to diagnose MS from the retinal optical coherence tomography (OCT) images. Recent clinical studies affirmed that thinning occurs on the four top layers, explicitly in the macular region. However, existing MS detection methods have not considered all MS symptoms, which may impact the MS detection performance. In this research, we propose a new automated method to detect MS from the retinal OCT images. The main principle is based on extracting the relevant retinal layers and figuring out the layer thicknesses, which are investigated to deduce the MS disease. The main challenge is to guarantee a higher performance biomarker extraction within an efficient exploration of OCT cuts. Our contribution consists of the following: (1) employing two DL architectures to segment separately sub-images based on their morphology, in order to enhance segmentation quality; (2) extracting thickness features from the four top layers; (3) dedicating a classifier for each OCT cut that is selected based on its position with respect to the macula center; and (4) merging the classifier knowledge through an ensemble learning approach. Our suggested method achieved 97% accuracy, 100% sensitivity, and 94% precision and specificity, which outperforms several state-of-the-art methods.

Accurate segmentation of lung adenocarcinoma nodules in computed tomography (CT) images is critical for clinical staging and diagnosis. However, irregular nodule shapes and ambiguous boundaries pose significant challenges for existing methods. This study introduces S³TU-Net, a hybrid CNN-Transformer architecture designed to enhance feature extraction, fusion, and global context modeling. The model integrates three key innovations: (1) structured convolution blocks (DWF-Conv/D²BR-Conv) for multi-scale feature extraction and overfitting mitigation; (2) S²-MLP Link, a spatial-shift-enhanced skip-connection module to improve multi-level feature fusion; and 3) residual-based superpixel vision transformer (RM-SViT) to capture long-range dependencies efficiently. Evaluated on the LIDC-IDRI dataset, S³TU-Net achieves a Dice score of 89.04%, precision of 90.73%, and IoU of 90.70%, outperforming recent methods by 4.52% in Dice. Validation on the EPDB dataset further confirms its generalizability (Dice, 86.40%). This work contributes to bridging the gap between local feature sensitivity and global context awareness by integrating structured convolutions and superpixel-based transformers, offering a robust tool for clinical decision support.

Accurate segmentation of hard exudate in fundus images is crucial for early diagnosis of retinal diseases. However, hard exudate segmentation is still a challenge task for accurately detecting small lesions and precisely locating the boundaries of ambiguous lesions. In this paper, the longitudinal multi-scale fusion network (LMSF-Net) is proposed for accurate hard exudate segmentation in fundus images. In this network, an adjacent complementary correction module (ACCM) is proposed on the encoding path for complementary fusion between adjacent encoding features, and a progressive iterative fusion module (PIFM) is designed on the decoding path for fusion between adjacent decoding features. Furthermore, a spatial awareness fusion module (SAFM) is proposed at the end of the decoding path for calibration and aggregation of the two decoding outputs. The proposed method can improve segmentation results of hard exudates with different scales and shapes. The experimental results confirm the superiority of the proposed method for hard exudate segmentation with AUPR of 0.6954, 0.9017, and 0.6745 on the DDR, IDRID, and E-Ophtha EX datasets, respectively.

Reliable feedback of gait variables, such as joint moments, is critical for designing controllers of intelligent assistive devices that can assist the wearer outdoors. To estimate lower extremity joint moments quickly and accurately outside the laboratory, a novel multimodal motion intent recognition system by fusing traditional deep learning models is proposed in this paper. The developed estimation method uses the joint kinematics data and individual feature parameters to estimate lower limb joint moments in the sagittal plane under different motion conditions: walking, running, and stair ascent and descent. Specifically, seven deep learning models that use combination of convolutional neural network, recurrent neural networks and attention mechanisms as the unit models of the framework are designed. To improve the performance of the unit models, a data augmentation module is designed in the system. Using those unit models, a novel framework, DeepMPSF-Net, which treats the output of each unit model as a pseudo-sensor observation and utilizes variable weight fusion methods to improve classification accuracy and kinetics estimation performance, is proposed. The results show that the augmented DeepMPSF-Net can accurately identify the locomotion, and the estimation performance (PCC) of joint moments is improved to 0.952 (walking), 0.988 (running), 0.925 (stair ascent), and 0.921 (stair descent), respectively. It also suggests that the estimation system is expected to contribute to the development of intelligent assistive devices for the lower limbs.

Monitoring user weight, including body weight and afforded load, is crucial for post-fracture rehabilitation. Inappropriate weight levels can delay recovery and increase re-fracture risk. In recent years, insole sensor systems have proven effective in monitoring gait parameters, including plantar pressure and gait cycles. Among all gait parameters, plantar pressure is particularly useful for monitoring and predicting user weight due to its strong correlation. However, previous studies were limited in scenarios and accuracy. To address these issues, this study proposes a piezoresistive plantar pressure sensor system (PPS) integrated with a CNN model. The system uses 96 piezoresistive force sensors to collect plantar pressure data from 107 subjects in both walking and standing conditions with varying loads (0 kg, 5 kg, 10 kg, 15 kg). The data is input into the CNN model for user weight prediction. Results show standing without load achieves an R² of 0.9997 and relative error of 0.0027, while walking with load shows the lowest R² of 0.8857 and relative error of 0.0416. This work enables accurate user weight estimation and supports gait-based healthcare research, particularly in relation to plantar pressure.

In today's world, rapid advancements in wireless sensor network (WSN) technologies hold the potential to revolutionize healthcare through future ubiquitous patient monitoring systems. Essential for continuous monitoring without restricting patient mobility, these systems comprise wearable or implanted sensors continuously tracking physiological parameters. Enabling seamless patient-doctor interaction, they monitor and transmit patient physiological data. This project involves designing an ECG monitoring system utilizing DigiMesh technology for wireless transmission to a remote device. Patient data is stored in the IoT-cloud via a MySQL database, enabling real-time remote monitoring by medical staff. The sensor node processes ECG data, transmitted to the Sink Node, and the MySQL database facilitates data storage. Utilizing a web-based system accessible on all devices, the proposed monitoring system displays ECG results, reports, and patient information. The goal is to create a reliable, cost-effective, low-power vital signs monitoring system transmitting various body parameters wirelessly to medical professionals. In hospitals, continuous monitoring is crucial for patients requiring extended medical care, ensuring constant surveillance even in non-emergency situations.

Consciousness assessment in disorders of consciousness (DoC) patients remains clinically challenging. Dynamic brain activities responsive to sensory stimulations have been suggested to contain consciousness-related information. However, primary sensory processing can occur unconsciously, necessitating evaluation of residual higher-order cognitive functions for effective assessment. In this study, we introduced a movie-viewing paradigm incorporating a scrambled version to control for primary sensory processing and applied electroencephalography (EEG) microstate analysis to capture higher-order neural dynamics. By comparing 23 DoC patients with 23 healthy individuals and 12 conscious brain-injured patients, we found significant abnormalities in microstate D in DoC patients. Healthy individuals and conscious brain-injured patients showed enhanced D-related parameters during intact movie-viewing compared to the scrambled condition. Conversely, DoC patients displayed a significant decrease in Duration, Coverage, Occurrence, and Transition Probabilities of microstate D during intact movie-viewing. Additionally, K-nearest neighbors classifier showed that the differences in microstate features between the intact and scrambled movie-viewing yielded the best classification outcome (AUC = 0.83), in which microstate D parameters serve as the most important features. Our results suggested that EEG microstates during naturalistic movie-viewing, especially microstate D, have the potential to serve as a novel, objective indicator for characterizing and diagnosing the state of consciousness.

In this paper, we propose a new orthodontic wire design system (OWDS) that allows medical staff to set the bracket attachment position and direction on a 3D tomographic medical image. To enable fully automated processing of the orthodontic wire by a robot, a method for modeling the geometrically designed wire based on homogeneous transformation is proposed. A new custom algorithm is proposed for optimal wire design, which results in the shortest length that satisfies the constraints required for wire mounting. Through case studies of wire geometry design and other numerical experiments, the effectiveness of the proposed method is verified.

Professional drivers are frequently exposed to multi-axis vibration environments. The effect of active control of neck muscles on injury risk of cervical spine in a sustained vibration environment remains unclear. This study aims to explore the effect of multi-axis vibration and muscles on driver's neck health. A head-neck finite element model with active muscles was developed. The injury of the cervical spine tissues was analyzed under different muscle activation levels and vibration conditions. The results showed that compared with the uniaxial vibration, the vibration amplitude of head increased by 70.65% (vertical) under the multi-axis vibration condition. This indicated that the vibration of the head-neck was more intense under the multi-axis vibration environment compared with the uniaxial vertical vibration. Compared with passive muscles, the active muscles (activation level was 0.05) could reduce the head vibration amplitude by 39.88% in the vertical direction, 12.59% in the fore-aft direction, respectively, and this vibration suppression was more pronounced in the fore-aft direction compared with the vertical direction. In the multi-axis vibration environment, neck muscles could suppress head movements induced by vibration, especially in the fore-aft direction, reducing the risk of disc injury. The level of muscle activation was positively correlated with the suppressive effect.

Early detection of adverse drug reactions (ADRs) is crucial for patient safety but remains challenging due to underreporting and delayed data in traditional pharmacovigilance. This study proposes a computationally efficient and interpretable framework for ADR detection by integrating Low-Rank Adaptation (LoRA) and SHapley Additive Explanations (SHAP) with encoder-based transformer models (BERT, DistilBERT, RoBERTa). Leveraging over 3,900 annotated tweets, our approach demonstrates that LoRA reduces trainable parameters and training costs by up to 50%, while preserving high classification accuracy (above 98%) across three disease classes. SHAP analysis provides actionable interpretability, revealing that the models consistently rely on clinically relevant terms, such as drug names and symptoms, to drive predictions. Compared to traditional finetuning, LoRA and Efficient Finetuning of Quantized LLMs (QLoRA) offer a robust and scalable alternative for processing noisy, informal social media data, making real-time ADR monitoring feasible in resource-constrained healthcare settings. This framework strikes a balance between computational efficiency, interpretability, and predictive performance, supporting the integration of pharmacovigilance into clinical decision support systems for safer patient care.