A5066196985- Image and Signal Denoising Methods
- Sparse and Compressive Sensing Techniques
- Mathematical Dynamics and Fractals
- Advanced Image Processing Techniques
- Blind Source Separation Techniques
- Advanced Image Fusion Techniques
- Machine Fault Diagnosis Techniques
- semigroups and automata theory
- Higher Education and Teaching Methods
- Formal Methods in Verification
- Medical Image Segmentation Techniques
- Metaheuristic Optimization Algorithms Research
- Advanced Data Compression Techniques
- Evolutionary Algorithms and Applications
- Image Retrieval and Classification Techniques
- Advanced Multi-Objective Optimization Algorithms
- Photoacoustic and Ultrasonic Imaging
- Advanced Neural Network Applications
- Advanced Topology and Set Theory
- Image Processing Techniques and Applications
- Computer Graphics and Visualization Techniques
- Petri Nets in System Modeling
- Domain Adaptation and Few-Shot Learning
- Advanced Image and Video Retrieval Techniques
- Medical Imaging Techniques and Applications
Shenzhen University
2016-2025
Dalian University
2018-2025
National Supercomputing Center in Shenzhen
2022-2024
Chinese Academy of Medical Sciences & Peking Union Medical College
2022-2024
Chinese People's Liberation Army
2024
Tongji University
2022
Guangdong Institute of Intelligent Manufacturing
2021
Syracuse University
2020
Bridge University
2020
Nanjing University
2002-2019
Wavelet transforms are multiresolution decompositions that can be used to analyze signals and images. They describe a signal by the power at each scale position. Edges located very effectively in wavelet transform domain. A spatially selective noise filtration technique based on direct spatial correlation of several adjacent scales is introduced. high infer there significant feature position should passed through filter. The authors have tested simulated signals, phantom images, real MR It...
Experience suggests the existence of a connection between contrast gray-scale image and gradient magnitude intensity edges in neighborhood where is measured. This observation motivates development edge-based enhancement techniques. We present simple effective method for based on multiscale edge representation images. The an can be enhanced simply by stretching or upscaling maxima image. offers flexibility to selectively enhance features different sizes ability control noise magnification....
Regular expression (regex) with modern extensions is one of the most popular string processing tools. However, poorly-designed regexes can yield exponentially many matching steps, and lead to regex Denial-of-Service (ReDoS) attacks under well-conceived inputs. This paper presents Rescue, a three-phase gray-box analytical technique, automatically generate ReDoS strings highlight vulnerabilities given regexes. Rescue systematically seeds (by genetic search), incubates another finally pumps...
Convolutional neural networks (CNNs) have been widely used in image super-resolution (SR). Most existing CNN-based methods focus on achieving better performance by designing deeper/wider networks, while suffering from heavy computational cost problem, thus hindering the deployment of such models mobile devices with limited resources. To relieve we propose a novel and efficient SR model, named Feature Affinity-based Knowledge Distillation (FAKD), transferring structural knowledge teacher...
Convolutional neural networks (CNNs) have come to dominate vision-based deep network structures in both image and video models over the past decade. However, convolution-free vision Transformers (ViTs) recently outperformed CNN-based recognition. Despite this progress, building designing not yet obtained same attention research as image-based Transformers. While there been attempts build by adapting for understanding, these still lack efficiency due large gap between regarding number of...
In recent years, latent diffusion models (LDMs) have successfully demonstrated their strong generative and generalization capabilities in the field of image generation. However, when applied to generate multiple-view same object, LDMs are unable ensure 3D-consistencies among generated multiple images. Existing generation methods leverages power CLIP embedding combined with camera produce semantically coherent such semantic multi-view consistent heuristics is rather weak does not necessarily...
Last decade witnesses an impressive development of embedded reactive systems, which motivates the re- search open where multiple components in- teract with each other and their environment these interactions decide behavior system. A natu- ral "common-denominator" model for systems is concurrent game structure, in several players can concurrently on Alternating-time temporal logic (ATL), a geared towards specification verification properties allows to reason existence strategies coalitions...
Recently, there has been significant interest in robust fractal image coding for the purpose of robustness against outliers. However, known methods (HFIC and LAD-FIC, etc.) are not optimal, since, besides high computational cost, they use corrupted domain block as independent variable regression model, which may adversely affect estimator to calculate parameters (depending on noise level). This paper presents a Huber fitting plane-based (HFPFIC) method. method builds planes (HFPs) range...