Currently, Coronavirus Disease 2019 (COVID-19) is still endangering world health and safety deep learning (DL) expected to be the most powerful method for efficient detection of COVID-19. However, patients' privacy concerns prohibit data sharing between medical institutions, leading unexpected performance neural network (DNN) models. Fortunately, federated (FL), as a novel paradigm, allows participating clients collaboratively train models without exposing source outside original location....

ABSTRACT The best screening method for detecting heteroresistant vancomycin-intermediate Staphylococcus aureus (hVISA) remains unclear. Using population analysis profiling utilizing the area under concentration-time curve (PAP-AUC) as gold standard, we screened 458 consecutive methicillin-resistant S. (MRSA) bloodstream isolates to determine most accurate and cost-effective testing strategy detect presence of heteroresistance. All were also tested using macromethod Etest (MET) glycopeptide...

Coronary heart disease is the first killer of human health. At present, most widely used approach coronary diagnosis angiography, a surgery that could potentially cause some physical damage to patients, together with complications and adverse reactions. Furthermore, angiography expensive thus cannot be in under development country. On other hand, color Doppler echocardiography report, blood biochemical indicators personal information, such as gender, age diabetes, can reflect degree patients...

This research aimed to assess the value of radiomics combined with multiple machine learning algorithms in diagnosis pancreatic ductal adenocarcinoma (PDAC) lymph node (LN) metastasis, which is expected provide clinical treatment strategies.

Without interrupting the traffic on old bridge, connection of widening bridge will cause disturbance to concrete in splicing position. In order study anti-disturbance performance material, Normal Concrete (NC) material and Ultra-High-Performance (UHPC) were experimentally investigated by vertical shaking table X-ray computed tomography scanning. It can be learnt from tests that compressive strength NC UHPC increased about 10% 20% after disturbance, while flexural tensile splitting both...

XML has become a de facto standard for data representation and exchange over the Internet. With emergence of more documents, clustering documents an active research area. lie between structured unstructured which describe both content structure, so how to effectively cluster is huge challenge. However, most existing algorithms are based on structural similarities not or less take into account documents. In this paper, we develop novel method measuring combines structure contents Based...

Research suggests that individuals who experience prolonged exposure to stress may be at higher risk for developing psychological disorders. Currently, is primarily evaluated by professional physicians using rating scales, which prone subjective biases and limitations of the scales. Therefore, it imperative explore more objective, accurate, efficient biomarkers evaluating level in an individual. In this study, we utilized 4D data-independent acquisition (4D-DIA) proteomics quantitative...

Background: To predict mycophenolic acid (MPA) exposure in renal transplant recipients using a deep learning model based on convolutional neural network with bilateral long short-term memory and attention methods. Methods: A total of 172 Chinese patients were enrolled this study. The divided into training group (n = 138, Ruijin Hospital) validation 34, Zhongshan Hospital). Fourteen days after transplantation, rich blood samples collected 0–12 hours MPA administration. plasma concentration...

Precise breast cancer classification can aid the clinicians in making decision of diagnosis and therapy. The existing works mainly utilized medical images for classification, ignoring value Electronic Medical Records (EMR). In this study, we investigate problem with EMR data propose a new model based on Hierarchical Attention Bidirectional Networks (HA-BiRNN). Our proposed supports integration features reports context information EMR. Specially, our first uses three HA-BiRNNs to extract from...

Diabetic complications have brought a tremendous burden for diabetic patients, but the problem of predicting is still unresolved. Our aim to explore relationship between hemoglobin A1C (HbA1c), insulin (INS), and glucose (GLU) in combination with individual factors effectively predict multiple diabetes.This was real-world study. Data were collected from 40,913 participants an average age 48 years Department Endocrinology Ruijin Hospital Shanghai. We proposed deep personal multitask...

We design and propose a compact street lamp based on dual-module chip-on-board LED. The is composed of six faceted reflectors. It can direct the luminous flux form uniform illumination target area, it effectively reduces power consumption. have conducted both simulations prototype measurements. test results show good optical performance in that uniformity luminance reaches 0.58 for LED zigzag arrangements 0.60 double-side arrangements. average fulfill requirements Chinese road lighting...

To explore the value of four-dimensional CT angiography (4D-CTA) in preoperative evaluation juvenile nasopharyngeal angiofibromas (JNAs) using 320-row volume CT.4D-CTA and DSA data 18 patients with histopathologically proven JNAs were retrospectively reviewed. The location, extent, feeding vessels stage assessed by two radiologists independently blindly. agreements between both reviewers 4D-CTA surgical findings for assessing above indicators analysed, respectively. radiation dose number...

Alzheimer's Disease (AD) is an irreversible neurodegenerative disease that commonly occurs in the elderly. With current accelerated aging process, accurate diagnosis of early AD essential for patient care and delay. In recent years, Magnetic Resonance Imaging (MRI) has become increasingly important diagnosing due to advances deep learning neuroimaging technology. This paper proposes a model framework multi-classification prediction based on fusing multi-modal features. Firstly, sMRI data are...

Deep neural networks(DNN)have achieved good results in the application of Named Entity Recognition (NER), but most DNN methods are based on large numbers annotated data. Electronic Medical Record (EMR) belongs to text data specific professional field. The annotation this kind needs experts with strong knowledge medical field and time labeling. To tackle problems areas, volume, difficulties EMR, we propose a new method multi-standard active learning recognize entities EMR. Our approach uses...