A5101668326- Advanced Neural Network Applications
- Medical Image Segmentation Techniques
- Total Knee Arthroplasty Outcomes
- Autonomous Vehicle Technology and Safety
- Video Surveillance and Tracking Methods
- Knee injuries and reconstruction techniques
- Maritime Navigation and Safety
- Wind Turbine Control Systems
- Advanced Image Fusion Techniques
- Fire Detection and Safety Systems
- Bayesian Modeling and Causal Inference
- HVDC Systems and Fault Protection
- Advanced Aircraft Design and Technologies
- Electric and Hybrid Vehicle Technologies
- Ovarian function and disorders
- Target Tracking and Data Fusion in Sensor Networks
- Traffic and Road Safety
- Radiomics and Machine Learning in Medical Imaging
- Control and Dynamics of Mobile Robots
- Data Quality and Management
- Advanced Battery Technologies Research
- Ammonia Synthesis and Nitrogen Reduction
- Cardiovascular Disease and Adiposity
- ECG Monitoring and Analysis
- Software Engineering Research
University of California, Riverside
2023
Affiliated Hospital of Southwest Medical University
2023
First Hospital of Shijiazhuang
2023
Hefei University of Technology
2012-2022
Shanghai Jiao Tong University
2022
Lanzhou University
2021-2022
Lanzhou University Second Hospital
2021-2022
Nanjing Medical University
2020
Abstract: Automated electrocardiogram (ECG) diagnosis could be a useful aid for clinical use. We applied deep learning method to build system automated detection and classification of ECG signals. first trained convolutional neural network (CNN) detect cardiovascular disease in signals using training data set 259,789 collected from the cardiac function rooms tertiary care hospital. The CNN was validated an independent test 18,018 labels used covered >90% diagnoses. grouped ECGs into 18...
Real-world programs expecting structured inputs often has a format-parsing stage gating the deeper program space. Neither mutation-based approach nor generative can provide solution that is effective and scalable. Large language models (LLM) pre-trained with an enormous amount of natural corpus have proved to be for understanding implicit format syntax generating format-conforming inputs. In this paper, propose ChatFuzz, greybox fuzzer augmented by AI. More specifically, we pick seed in...
Superpixel is generated by automatically clustering pixels in an image into hundreds of compact partitions, which widely used to perceive the object contours for its excel-lent contour adherence. Although some works use Convolution Neural Network (CNN) generate high-quality superpixel, we challenge design principles these net-works, specifically their dependence on manual labels and excess computation resources, limits flexibility compared with traditional unsupervised segmentation methods....
To minimize the killer turn caused by sharp margin of tibial tunnel exit in transtibial PCL reconstruction, surgeons tend to maximize angle relation plateau. However, date, no consensus has been reached regarding maximum for tunnel.In this study we sought (1) determine anteromedial and anterolateral approaches reconstruction; (2) compare differences based on three measurement methods: virtual radiographs, CT images, three-dimensional (3D) knee models; (3) conduct a correlation analysis...
In this paper, energy management strategy (EMS) model based on deep recurrent neural network (DRNN) is presented to learn optimal torque distribution for the single-axle parallel hybrid electric vehicle. The has two distinguishing properties: 1) because EMS formulated as a time series prediction problem, taking historical data input of captures input-and-output dynamic characteristics and enhances capability 2) end-to-end framework directly generates results without extracting features...
Objective The administration of progesterone before transfer in hormone replacement treatment (HRT) is crucial for the clinical outcomes frozen-thawed embryo (FET), but optimal duration remains controversial. This study aimed to investigate effect on FET cycles. Methods prospective cohort included 353 artificial cycles conducted at a reproductive medicine center between April and October 2021. were stratified into four groups based supplementation procedure embryonic development stage: group...
A motion planning method that integrates global and local control is proposed in this paper, aiming to enable Automated Guided Vehicle (AGV) navigate automatically two-dimensional environment. The layered cost maps are utilized divided into map respectively provide environment information for the control. Specifically, based on path algorithm establish a optimal map. Under guidance of path, generates an collision free trajectory with temporal map, which can avoid contact surrounding...
Currently, the analysis of cooling, heating and power (CCHP) system is usually based on operation strategy (such as following electric load) optimization criterion cost reduction) without considering match between sources loads distributions from perspective allocation proportion waste heat cooling units. And this issue taken into consideration in paper. An objective function minimum proposed, along with constrains capacity parameters CCHP system, including four different strategies. This...
Motion decision-making is an open-challenging issue for autonomous driving, especially in the complex and diverse environment. A motion model based on deep reinforcement learning (DRL) proposed this work. To optimize driving policy, a multi-objective reward function designed to guide system explore optimal decision policy with goal of safety, efficiency, smoothness. convolutional neural network (CNN) as backbone DRL make full use observation information enhance safety check module avoiding...
Abstract Semantic segmentation is crucial to the autonomous driving, as an accurate recognition and location of surrounding scenes can be provided for street understanding task. Many existing networks usually fuse high‐level low‐level features boost performance. However, simple fusion may impose a limited performance improvement because gap between features. To alleviate this limitation, we respectively propose spatial aggregation channel bridge gap. Our implementation, inspired by attention...
With increasing penetration of renewable energy generators, such as wind doubly-fed induction generators (DFIGs), semiconductor based power converters are getting widespread applications into the traditional system. Under this trend, future smart grid will become definitely a electronic dominated This change and evolution make it hard to describe interactions between system in electromechanical or electromagnetic transient time-scale. Moreover, generation systems including often deployed...
Knowledge unit (KU) is the smallest integral knowledge object in a given domain. relation recognition to discover implicit relations among KUs, which crucial problem information extraction. This paper proposes framework based on Markov Logic Networks, combines probabilistic graphical models and first-order logic by attaching weight each formula. The composed principally of structure learning, artificial add or delete formulas, learning inferring. According semantic analysis KUs their...
Abstract Background Interference screw is commonly used for graft fixation in anterior cruciate ligament (ACL) reconstruction. However, previous studies had reported that the insertion of interference screws significantly caused laceration. The purposes this study were to (1) quantitatively evaluate laceration from one single PEEK screws; and (2) determine whether different types sutures reduced after ACL Methods in-vitro reconstruction model was created using porcine tibias bovine extensor...
Graph contrastive learning is a general paradigm excelling at capturing invariant information from diverse perturbations in graphs. Recent works focus on exploring the structural rationale graphs, thereby increasing discriminability of information. However, such methods may incur mis-learning graph models towards interpretability and thus learned noisy task-agnostic interferes with prediction To this end, purpose intrinsic we accordingly propose to capture dimensional which has not received...
Objective To study the effect of extracellular matrix(ECM)of xenogenic femoral fascia which is a tissue-engineering material in repair renal trauma.Methods Twenty-four experiment dogs were divided into 3 groups:group 1(n=10),the kidneys repaired using ECM fascia;group 2(n=10),the self-omentum;group 3(n=4),xenogenic was used as materials.The animals sacrificed separately at 1,2 weeks and 1,2,4 months after operations group 1,2.In 3,the 2 operations.The examinations blood routine performed...
Objective To evaluate the ability of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) in identifying rmpA2-carrying hypervirulentKlebsiella pneumoniae. Methods A total 57nonduplicateKlebsiella pneumoniae isolates were collected from Second Affiliated Hospital Zhejiang University and Henan Provincial People′s Hospital. Virulence gene rmpA2 capsule K serotype-specific genes detected by PCR, multilocus sequence typing (MLST) was performed for...
Aiming at the problems of poor robustness and single low accuracy sensor in obstacle detection, this paper proposes multi-source information fusion method based on extended Kalman filter for position data radar LiDAR. By means coordinate transformation, collected by two radars are unified to polar system, is constructed according state vector data, specific parameters selected fuse location sensors. The simulation results show that algorithm can effectively different sensors improve...
Superpixel is generated by automatically clustering pixels in an image into hundreds of compact partitions, which widely used to perceive the object contours for its excellent contour adherence. Although some works use Convolution Neural Network (CNN) generate high-quality superpixel, we challenge design principles these networks, specifically their dependence on manual labels and excess computation resources, limits flexibility compared with traditional unsupervised segmentation methods. We...