A5101868894- Advanced Neural Network Applications
- Domain Adaptation and Few-Shot Learning
- Advanced Image and Video Retrieval Techniques
- Psychosocial Factors Impacting Youth
- Industrial Vision Systems and Defect Detection
- Emotional Labor in Professions
- Intergenerational Family Dynamics and Caregiving
- Internet Traffic Analysis and Secure E-voting
- Health and Well-being Studies
- Privacy-Preserving Technologies in Data
- Resilience and Mental Health
- Vehicle License Plate Recognition
- Video Surveillance and Tracking Methods
- Cancer Mechanisms and Therapy
- Impact of Light on Environment and Health
- PI3K/AKT/mTOR signaling in cancer
- Regional Development and Environment
- Network Security and Intrusion Detection
- Fire Detection and Safety Systems
- Housing Market and Economics
- International Arbitration and Investment Law
- Simulation and Modeling Applications
- Housing, Finance, and Neoliberalism
- Machine Learning and Data Classification
- Business Process Modeling and Analysis
China Academy of Engineering Physics
2021-2024
Command Hospital
2024
Tongji University
2019-2023
National University of Defense Technology
2023
Affiliated Hospital of Southwest Medical University
2023
China Mobile (China)
2023
China Three Gorges University
2019
Nanjing University of Posts and Telecommunications
2019
Jining Medical University
2018
Peking University
2004
Federated learning is a novel framework that enables resource-constrained edge devices to jointly learn model, which solves the problem of data protection and islands. However, standard federated vulnerable Byzantine attacks, will cause global model be manipulated by attacker or fail converge. On non-iid data, current methods are not effective in defensing against attacks. In this paper, we propose Byzantine-robust for via credibility assessment on (BRCA). Credibility designed detect attacks...
Recognizing speed limit information is crucial for advanced driver assistance systems (ADAS) as it directly affects the safety planning and decision-making process of intelligent driving systems. However, traditional image recognition-based solutions confront inherent restrictions precision issues due to uncontrolled external factors. This paper endeavors present a novel, data-driven solution recognition that leverages stability maturity technologies, overcoming these challenges. We...
<title>Abstract</title> <bold>Background and objective</bold>: The correlation of cardiopulmonary exercise testing (CPET) parameters the prognosis coronary artery disease (CAD) patients with high pulse pressure (PP) is uncertain. Present study evaluated association value CPET in PP patients. <bold>Methods</bold>: Patients CAD who underwent percutaneous intervention (PCI) were enrolled. Enrolled divided into two groups according to after admission: group normal group. primary endpoint was...
Abstract Objective . The primary purpose of this work is to demonstrate the feasibility a deep convolutional neural network (dCNN) based algorithm that uses two-dimensional (2D) electronic portal imaging device (EPID) images and CT as input reconstruct 3D dose distributions inside patient. Approach To generalize dCNN training testing data, geometric materials models VitalBeam accelerator treatment head corresponding EPID imager were constructed in detail GPU-accelerated Monte Carlo computing...
Capsule network was introduced as a new architecture of neural networks, it encoding features capsules to overcome the lacking equivariant in convolutional networks. It uses dynamic routing algorithm train parameters different capsule layers, but need be improved. In this paper, we propose novel and discussed effect initialization method coupling coefficient $c_{ij}$ on model. First, analyze rate change initial value when iterates. The larger $c_{ij}$, better Then, proposed improvement that...
Capsule network is the most recent exciting advancement in deep learning field and represents positional information by stacking features into vectors. The dynamic routing algorithm used capsule network, however, there are some disadvantages such as inability to stack multiple layers a large amount of computation. In this paper, we propose an adaptive that can solve problems mentioned above. First, low-layer capsules adaptively adjust their direction length removing influence coupling...
Since it is difficult to get texture details for sonar image denoising with strong noise and weak feature information. A enhancement algorithm in wavelet domain based on Gaussian mixture model proposed preserve the information of image. Under interference, traditional method has certain difficulties measuring similarity For this reason, multi-scale analysis performed extract each resolution. Secondly, a directed probability map between adjacent scales constructed realize similar association....
Purple sweet potato is rich in anthocyanins and has a very high use value.The method of extracting anthocyanin from purple analyzed, various extraction methods are compared detail, the existing problems purification proposed, future development direction technology pointed out.
As an important research direction of computer vision, panoramic mosaic technology has a very position in many fields such as disaster detecting and terrain reconnaissance, target detection tracking, drone aerial satellite survey, virtual reality so on. Although researchers have made lot achievements the field stitching, it is found that its universality real-time performance need to be improved practical engineering applications. In order solve problems existing algorithms poor when feature...