A5103160972- Advanced SAR Imaging Techniques
- Radar Systems and Signal Processing
- Sparse and Compressive Sensing Techniques
- Synthetic Aperture Radar (SAR) Applications and Techniques
- Distributed Sensor Networks and Detection Algorithms
- Blind Source Separation Techniques
- Remote-Sensing Image Classification
- Statistical Methods and Inference
- Neural Networks and Applications
- Information and Cyber Security
- Fault Detection and Control Systems
- Photoacoustic and Ultrasonic Imaging
- Smart Grid and Power Systems
- Natural Language Processing Techniques
- Advanced Algorithms and Applications
- Software System Performance and Reliability
- Direction-of-Arrival Estimation Techniques
- Advanced Text Analysis Techniques
- Bayesian Methods and Mixture Models
- Microwave Imaging and Scattering Analysis
- Advanced Decision-Making Techniques
- Gaussian Processes and Bayesian Inference
- Structural Health Monitoring Techniques
- Imbalanced Data Classification Techniques
- Organic Chemistry Cycloaddition Reactions
Shenzhen Research Institute of Big Data
2022-2025
Hohai University
2024
Chongqing University of Posts and Telecommunications
2024
Chinese University of Hong Kong, Shenzhen
2022-2024
Ministry of Agriculture and Rural Affairs
2022
South China Agricultural University
2020-2022
Zhejiang Energy Research Institute
2021
Shanghai Power Equipment Research Institute
2016-2018
<title>Abstract</title> This study provides an in-depth analysis of the optimized design shear wall layout in high-rise buildings, aiming to improve seismic performance and reduce construction costs. Through experiments on a building model from residential community Jiangsu Province, results show that when intensity reaches level 7 (0.1g), traditional has ratio 4.24%, while optimal model's does not exceed under same conditions, demonstrating its flexibility adaptability. As increases...
Ship classification based on synthetic aperture radar (SAR) images is a crucial component in maritime surveillance. In this article, the feature selection and classifier design, as two key essential factors for traditional ship classification, are jointed together, novel model combining kernel extreme learning machine (KELM) dragonfly algorithm binary space (BDA), named BDA-KELM, proposed which conducts automatic searches optimal parameter sets (including penalty factor) at same time....
Nowadays it has still remained as a big challenge to efficiently compress color images in the encrypted domain. In this paper we present novel deep-learning-based approach encryption-then-lossy-compression (ETC) of by incorporating domain knowledge image reconstruction process. specific, simple yet effective uniform down-sampling is utilized for lossy compression with modulo-256 addition, and task from an down-sampled then formulated problem constrained super-resolution (SR) reconstruction....
Localizing the root cause of network faults is crucial to operation and maintenance (O&M). Significant operational expenses will be saved if can identified agilely accurately. However, this challenging for human beings due complicated wireless environments architectures. Resorting data analysis machine learning promising but remains difficult various practical issues, such as lack well-labeled samples, hybrid fault behaviors, missing data, so on. In paper, we introduce a novel real-world...
The problem of joint direction-of-arrival estimation and distorted sensor detection has received a lot attention in recent decades. Most state-of-the-art work formulated such via low-rank row-sparse decomposition, where the components were treated an isolated manner. Such formulation results performance loss. Differently, this paper, we entangle by exploring their inherent connection. Furthermore, take into account maximal distortion level sensors. An alternating optimization scheme is...
Integer linear programs (ILPs) are commonly employed to model diverse practical problems such as scheduling and planning. Recently, machine learning techniques have been utilized solve ILPs. A straightforward idea is train a via supervised learning, with an ILP the input optimal solution label. An symmetric if its variables can be permuted without changing problem structure, resulting in numerous equivalent solutions. Randomly selecting label introduce variability training data, which may...
The global energy crisis is increasingly severe, and the construction industry, as a high-energy-consuming sector, one of main sources carbon emissions. As result, development green buildings has become imperative. Shear walls, common structural form in buildings, have their wall layout ratio significantly influencing amount building materials used, which crucial for material reduction emission during construction. This paper innovatively introduces concept optimal shear ratio, focusing on...
Azimuth multichannel SAR system can realize high-resolution and wide-swath (HRWS) imaging, which has attracted intensive research interests. The azimuth ambiguities of be suppressed by digital beamforming (DBF) techniques. However, the presence inevitable channel errors deteriorates performance ambiguity suppression. In this paper, we focus on range-variant phase errors. We analyze effect radial position using second order Taylor series to expand history. It indicates that regarded as divide...
Linear regression is a simple but effective method for data fitting to implement prediction. Research of linear approach real-time power load prediction was introduced in this paper market participants. With model simplification and precision improvement application by stepwise AIC Cook's distance analysis improvement, the scheme aiming at designing easy reliable provide participants optimized trading decision support among day-ahead real time spot market.
The problem of joint direction-of-arrival estimation and distorted sensor detection has received a lot attention in recent decades. Most state-of-the-art work formulated such via low-rank row-sparse decomposition, where the components were treated an isolated manner. Such formulation results performance loss. Differently, this paper, we entangle by exploring their inherent connection. Furthermore, take into account maximal distortion level sensors. An alternating optimization scheme is...
In this article, the estimate of high-frequency phase error (HPE) induced by residual motion (RME) is studied for synthetic aperture radar (SAR). The proposed algorithm contains two basic procedures. First, according to pair-echoes caused HPE, frequency coefficients could be estimated mean filter. Second, amplitudes and initial phases accomplished parametric or nonparametric autofocus manners. sharpness metric (SM) used in manner, while notch filter (NF) gradient (PGA) are performed other...
For complex text data, especially for long in order to measure the similarity, traditional methods are not accurate enough. We found that it is mainly because feature representation ability strong To improve accuracy of an algorithm based on pre-training deep learning model proposed extract features text. On benchmark data set THUCNews corpus, our method 5.4% higher than algorithm. Besides, we perform ablation experiments test improvement fine-tuning technology.
Virtual tour among sparse 360$^\circ$ images is widely used while hindering smooth and immersive roaming experiences. The emergence of Neural Radiance Field (NeRF) has showcased significant progress in synthesizing novel views, unlocking the potential for scene exploration. Nevertheless, previous NeRF works primarily focused on object-centric scenarios, resulting noticeable performance degradation when applied to outward-facing large-scale scenes due limitations parameterization. To achieve...