A5101678738- Advanced Image Processing Techniques
- Image and Signal Denoising Methods
- Data Mining Algorithms and Applications
- Image Processing Techniques and Applications
- Rough Sets and Fuzzy Logic
- Advanced Vision and Imaging
- Advanced Computational Techniques and Applications
- Advanced Decision-Making Techniques
- ICT Impact and Policies
- Data Stream Mining Techniques
- Forensic Toxicology and Drug Analysis
- Advanced Image Fusion Techniques
- Advanced Optimization Algorithms Research
- Homotopy and Cohomology in Algebraic Topology
- Internet Traffic Analysis and Secure E-voting
- Software Testing and Debugging Techniques
- Software Reliability and Analysis Research
- Advanced Data Compression Techniques
- Algebraic structures and combinatorial models
- Network Security and Intrusion Detection
- Multi-Criteria Decision Making
- Auction Theory and Applications
- Big Data and Business Intelligence
- Open Source Software Innovations
- Advanced Topics in Algebra
Nankai University
2004-2022
Baidu (China)
2022
ETH Zurich
2022
Data Assurance and Communication Security
2022
Dalian University of Technology
2020
Nanjing Xiaozhuang University
2018
Institute of Information Engineering
2016
Chinese Academy of Sciences
2016
Shanghai Optical Instrument Research Institute
2011
Lanzhou University of Technology
2010
This paper reviews the NTIRE 2022 challenge on efficient single image super-resolution with focus proposed solutions and results. The task of was to super-resolve an input a magnification factor ×4 based pairs low corresponding high resolution images. aim design network for that achieved improvement efficiency measured according several metrics including runtime, parameters, FLOPs, activations, memory consumption while at least maintaining PSNR 29.00dB DIV2K validation set. IMDN is set as...
With the recently massive development in convolution neural networks, numerous lightweight CNN-based image super-resolution methods have been proposed for practical deployments on edge devices. However, most existing focus one specific aspect: network or loss design, which leads to difficulty of minimizing model size. To address issue, we conclude block devising, architecture searching, and design obtain a more efficient SR structure. In this paper, an edge-enhanced feature distillation...
Recent research on deep convolutional neural networks (CNNs) has provided a significant performance boost efficient super-resolution (SR) tasks by trading off the and applicability. However, most existing methods focus subtracting feature processing consumption to reduce parameters calculations without refining immediate features, which leads inadequate information in restoration. In this paper, we propose lightweight network termed DDistill-SR, significantly improves SR quality capturing...
ConvNets can compete with transformers in high-level tasks by exploiting larger receptive fields. To unleash the potential of ConvNet super-resolution, we propose a multi-scale attention network (MAN), coupling classical mechanism emerging large kernel attention. In particular, proposed (MLKA) and gated spatial unit (GSAU). Through our MLKA, modify gate schemes to obtain abundant map at various granularity levels, thereby aggregating global local information avoiding blocking artifacts....
Combinatorial testing (CT) seeks to detect potential faults caused by various interactions of factors that can influence the software systems. When applying CT, it is a common practice first generate set test cases cover each possible interaction and then identify failure-inducing after failure detected. Although this conventional procedure simple forthright, we conjecture not ideal choice in practice. This because 1) testers desire root cause failures before all needed are generated...
In this paper, four types of Levitin-Polyak well-posedness generalized vector equilibrium problems with both abstract set constraints and functional are investigated. Criteria characterizations for these obtained.
Data mining technology is a useful tool for knowledge discovery from large-scale databases. At present, most data researchers pay much attention to technique problems developing models and methods, while little basic issues of mining. In this paper, we address question propose domain-oriented data-driven acquisition model. A algorithms are also proposed show the validity model..
Recent years have witnessed significant advances in single image deblurring due to the increasing popularity of electronic imaging equipment. Most existing blind algorithms focus on designing distinctive priors for blur kernel estimation, which usually play regularization roles deconvolution formulation. However, little research effort has been devoted relative scale ambiguity between latent and kernel. The well-known L 1 normalization constraint, i.e., fixing sum all weights be one, is...
Decision Suppose System is playing an important role in computer science, technology and engineering, while intelligent decision-making one of the current hotspots. Intelligent methods their algorithms are most basics key cores information processing, pervasive computing so on. In view multi-attribute under linguistic setting, propose new decision method. Firstly construct a range pole plan introduce policy-maker risk-preference weight because attributes? measure-value uncertainty. Then with...
Recently, learned image compression techniques have achieved remarkable performance, even surpassing the best manually designed lossy coders. They are promising to be large-scale adopted. For sake of practicality, a thorough investigation architecture design compression, regarding both performance and running speed, is essential. In this paper, we first propose uneven channel-conditional adaptive coding, motivated by observation energy compaction in compression. Combining proposed grouping...