A5100449582- Image Retrieval and Classification Techniques
- Traffic Prediction and Management Techniques
- Advanced Sensor and Control Systems
- Synthetic Aperture Radar (SAR) Applications and Techniques
- Smart Grid and Power Systems
- Machine Fault Diagnosis Techniques
- Architecture and Computational Design
- Gear and Bearing Dynamics Analysis
- Art, Technology, and Culture
- Machine Learning and Data Classification
- Morphological variations and asymmetry
- Human Mobility and Location-Based Analysis
- Soft Robotics and Applications
- Advanced Vision and Imaging
- Remote-Sensing Image Classification
- Smart Materials for Construction
- Robotic Mechanisms and Dynamics
- Aerospace Engineering and Energy Systems
- Icing and De-icing Technologies
- Industrial Vision Systems and Defect Detection
- Advanced Algorithms and Applications
- Privacy-Preserving Technologies in Data
- Blockchain Technology Applications and Security
- Flow Measurement and Analysis
- Medical Image Segmentation Techniques
Central China Normal University
2024-2025
Hangzhou Dianzi University
2024
Northeast Forestry University
2024
Tianjin University of Technology
2019-2023
Beijing Institute of Big Data Research
2022
Nanjing University of Aeronautics and Astronautics
2022
Institute of Precision Mechanics
2021
Dartmouth College
2003
Dartmouth Hospital
2003
Modern privacy regulations have spurred the evolution of machine unlearning, a technique enabling trained model to efficiently forget specific training data. In prior unlearning methods, concept "data forgetting" is often interpreted and implemented as achieving zero classification accuracy on such Nevertheless, authentic aim achieve alignment between unlearned gold model, i.e., encouraging them identical accuracy. On other hand, exhibits non-zero due its generalization ability. To aligned...
WaveMLP has demonstrated remarkable performance in various vision tasks; such as dense feature detection and semantic segmentation. However, WaveMLP, a local model, imposes limitations on fully connected layers by only allowing connections between tokens within the same window. This constraint makes model neglect relationship among different windows, leading to token fusion degraded modelling performance. Specially, it poses challenges when dealing with hyperspectral image (HSI)...
Confocal microscopy is an essential component of wafer defect detection systems. Wafers are raw materials used in the manufacture semiconductor chips. The chip manufacturing process undergoes frequent updates, which cause increase number and types defects. This leads to lengthy scanning times for large wafers, warrants need enhance throughput optical inspections. To address this issue, we propose use multi-scale residual dilated convolution attention mechanism network (MRDCAN)...
Wind turbine icing often occurs in winter, which changes the aerodynamic characteristics of blades and reduces work efficiency wind turbine. In this paper, glaze ice model is established for horizontal-axis 3-D. The contains grid generation, two-phase simulation, heat mass transfer. Results show that smaller suffers from more serious problem, reflects on a larger thickness. Both collision transfer coefficient increase under size condition.
With the development of information technology, various social fields have gradually realised informatisation and digitisation, today's society has entered era big data. A large amount data exists in different departments systems. At present, cross-organisational sharing is often by point-to-point authorisation. However, there are still some threats management control schedule for whole region. To realise secure credible sharing, this paper proposes a data-sharing framework based on...
In this paper, the robot system of dual-arm cooperative assembly is designed and analyzed in view scene axle-hole industrial production. The d-h method used to analyze kinematics 6-dof robot, forward inverse equations are derived. Then image processed by algorithm reduce noise fluctuation interference, so as obtain more effective information facilitate accurate positioning visual system. This paper discusses how identify different color axes deep learning halcyon environment. Finally has...
Planetary gearboxes (PGs) are the core transmission components of complex rotating machinery. The diverse forms internal gear faults often complicate diagnosis and maintenance work. To improve fault accuracy PGs, this paper proposes a method for diagnosing PG based on deep autoencoders (DAE) convolutional neural networks (CNN). fully utilizes learning capability DAE combines it with precise recognition ability CNN features. Through in encoding layer, achieves dimensionality reduction...
The direct method is to compare the pixel difference between two frames of images. However, gray image and photometric error only keep convexity in a small range. Therefore, when there large displacement camera position, system may fall into suboptimal local minimum. Semantic features can solve this problem, but they be run scenarios where semantic are known. Line extract edge information have good for displacement. In addition, it easy adapt different scenes. letter, we propose line feature...
Abstract The drawing cushion of press line in the workshop whole car factory has been extensively upgraded from air to hydraulic cushion. is more accurate than force cushion, however, a complex mechanical, hydraulic, Motion Control and detection technology, coupled with press, which makes fault diagnosis difficult. This paper mainly introduces problems coordination among three systems electronic control Rexroth Company’s servo-hydraulic its intelligent method.