Remote photoplethysmography (rPPG) has recently attracted much attention due to its non-contact measurement convenience and great potential in health care computer vision applications. Early rPPG studies were mostly developed on self-collected uncompressed video data, which limited their application scenarios that require long-distance real-time transmission, also hindered the generation of large-scale publicly available benchmark datasets. In recent years, with popularization...

Stroke, as a disease with high incidence and mortality rates, is increasingly receiving attention. Despite the rapid development of deep learning in medical field providing excellent performance for AI-assisted diagnosis, automated segmentation stroke still poses significant challenges. Issues such similarity between hemorrhagic regions background, irregularity areas, vast variability hemorrhage sizes persist. To address these challenges, this paper introduces new network architecture that...

Cardiovascular disease has become one of the main diseases threatening human life and health. This is very common troublesome, existing medical resources are scarce, so it necessary to use a computer-aided automatic diagnosis overcome these limitations. A diagnostic system can automatically diagnose through an electrocardiogram (ECG) signal. paper proposes novel deep-learning method for ECG classification based on adversarial domain adaptation, which solves problem insufficient-labeled...

Extreme learning machine based on local receptive fields (ELM-LRFs) is a very fast method that can be used for feature extraction and classification. Bidirectional long-short time memory network (BLSTM), widely type of recurrent neural (RNN) architecture, has showed excellent performance in series processing fields. In this paper, we combine the superiority above algorithms propose accurate hybrid deep model which named DELM-LRF-BLSTM ECG signal recognition. This uses segmented heartbeats as...

Abstract Electrocardiogram (ECG) is mostly used for the clinical diagnosis of cardiac arrhythmia due to its simplicity, non-invasiveness, and reliability. Recently, many models based on deep neural networks have been applied automatic classification with great success. However, most independently extract internal features each lead in 12-lead ECG during training phase, resulting a lack inter-lead features. Here, we propose general model two-dimensional ResNet detached squeeze-and-excitation...

A gait energy image contains much information, which is one of the most effective means to recognize characteristics. The accuracy recognition greatly affected by covariates, such as viewing angle, occlusion clothing, and walking speed. Gait features differ somewhat angles. Therefore, how improve a cross-view challenging task. This study proposes new algorithm structure. Gabor filter used extract from images, since it can different directions scales. We use linear discriminant analysis (LDA)...

The goal of Clinical Named Entity Recognition (CNER) is to identify clinical terms from medical records, which great importance for subsequent research. Most the current Chinese CNER models use a single set features that do not consider linguistic characteristics language, e.g., they both word and character features, lack morphological information specialized lexical on characters in field. We propose RoBerta Glyce-Flat Lattice Transformer-CRF (RG-FLAT-CRF) model address this problem. uses...

The precise identification of arrhythmia is critical in electrocardiogram (ECG) research. Many automatic classification methods have been suggested so far. However, efficient and accurate still a challenge due to the limited feature extraction model generalization ability. We integrate attention mechanism residual skip connection into U-Net (RA-UNET); besides, between RA-UNET block executed as convolutional neural network (RA-CNN) for classification. was evaluated using MIT-BIH database...

QRS complex detection plays an important role in the identification and classification of ECG signals. In this paper, a new algorithm named II-P&T is proposed based on I-P&T by means IPCMM rather than fdatool to generate bandpass digital filter lowpass filter. Simulation demonstrates that MIT-BIH database, performance better P&T terms sensitivity (Se), positive predictivity (PP), error (DER) with results Se = 99.62%, PP 99.54%, DER 0.84%. The accuracy II-P&T+SVM reaches 97.68%, which higher...

Researchers have introduced side information such as social networks or knowledge graphs to alleviate the problems of data sparsity and cold starts in recommendation systems. However, most methods ignore exploration feature differentiation aspects propagation process. To solve above problem, we propose a new attention method based on an enhanced perception. Specifically, capture user preferences fine-grained manner graph, asymmetric semantic mechanism is adopted. It identifies influence...

Image segmentation is a crucial task in artificial intelligence fields such as computer vision and medical imaging. While convolutional neural networks (CNNs) have achieved notable success by learning representative features from large datasets, they often lack geometric priors global object information, limiting their accuracy complex scenarios. Variational methods like active contours provide theoretical interpretability but require manual initialization are sensitive to hyper-parameters....

A constraint-based sensor network nodes particle swarm search localization algorithm (CPL) is presented. First of all, a constraint domain an unknown node must be determined; Then the positions which meet specific criteria searched out by optimization and searching results within are recorded; Finally, node's can obtained calculating average recording results. As shown in experiment results, CPL has strong robustness, comparing with normal schemes such as least square method (LS), CPL's...

Abstract Electrocardiogram (ECG) is mostly used for clinical diagnosis of cardiac arrhythmia due to its simplicity, non-invasiveness, and reliability. Recently, many models based on the Deep Neural Networks (DNNs) have been applied automatic classification with great success. However, most independently extract internal features each lead in 12-lead ECG during training phase, resulting a lack inter-lead features. Here, we propose general model two-dimensional ResNet detached...

The pedometer combined with Bluetooth 4.0 can effectively implement the communication laptops, PAD? and mobile phones. it more accurately calculate number of pace through method real-time digital filter smoothing data curve, design not only enhance people's enthusiasm taking exercise but also save resource. Because technology is in 2.4 GHz band which free spectrum [1], bound to get widely applications recognition, put recorded by sent phone other modern terminal devices technology, then...