A5100336606- Advanced Image Processing Techniques
- Advanced Vision and Imaging
- Image and Signal Denoising Methods
- Image Processing Techniques and Applications
- Generative Adversarial Networks and Image Synthesis
- Image Enhancement Techniques
- Geochemistry and Geologic Mapping
- Optical Network Technologies
- Organic Light-Emitting Diodes Research
- Geological and Geochemical Analysis
- Advanced Memory and Neural Computing
- Computer Graphics and Visualization Techniques
- earthquake and tectonic studies
- Organic Electronics and Photovoltaics
- Medical Imaging Techniques and Applications
- Advanced Fiber Laser Technologies
- Advanced Measurement and Detection Methods
- Advanced Battery Materials and Technologies
- Advanced Image and Video Retrieval Techniques
- Machine Fault Diagnosis Techniques
- Advancements in Battery Materials
- Financial Risk and Volatility Modeling
- Advanced Mathematical Modeling in Engineering
- Smart Grid and Power Systems
- Luminescence and Fluorescent Materials
Northwestern Polytechnical University
2023-2025
Songshan Lake Materials Laboratory
2023-2025
Huawei Technologies (Sweden)
2024
Shenyang Aerospace University
2024
Qiqihar University
2024
Chinese University of Hong Kong
2020-2024
Xi'an University of Technology
2024
Chongqing University of Technology
2023
Beijing Institute of Technology
2023
University of Hong Kong
2023
Recent studies have shown the importance of modeling long-range interactions in inpainting problem. To achieve this goal, existing approaches exploit either standalone attention techniques or transformers, but usually under a low resolution consideration computational cost. In paper, we present novel transformer-based model for large hole inpainting, which unifies merits transformers and convolutions to efficiently process high-resolution images. We carefully design each component our...
Reference-based image super-resolution (RefSR) has shown promising success in recovering high-frequency details by utilizing an external reference (Ref). In this task, texture are transferred from the Ref to low-resolution (LR) according their point- or patch-wise correspondence. Therefore, high-quality correspondence matching is critical. It also desired be computationally efficient. Besides, existing RefSR methods tend ignore potential large disparity distributions between LR and images,...
Single image super-resolution (SISR) deals with a fundamental problem of upsampling low-resolution (LR) to its high-resolution (HR) version. Last few years have witnessed impressive progress propelled by deep learning methods. However, one critical challenge faced existing methods is strike sweet spot model complexity and resulting SISR quality. This paper addresses this pain point proposing linearly-assembled pixel-adaptive regression network (LAPAR), which casts the direct LR HR mapping...
We consider the single image super-resolution (SISR) problem, where a high-resolution (HR) is generated based on low-resolution (LR) input. Recently, generative adversarial networks (GANs) become popular to hallucinate details. Most methods along this line rely predefined single-LR-single-HR mapping, which not flexible enough for ill-posed SISR task. Also, GAN-generated fake details may often undermine realism of whole image. address these issues by proposing best-buddy GANs (Beby-GAN)...
Pre-training has marked numerous state of the arts in high-level computer vision, while few attempts have ever been made to investigate how pre-training acts image processing systems. In this paper, we tailor transformer-based regimes that boost various low-level tasks. To comprehensively diagnose influence pre-training, design a whole set principled evaluation tools uncover its effects on internal representations. The observations demonstrate plays strikingly different roles For example,...
In this work, we focus on synthesizing high-quality textures 3D meshes. We present Point-UV diffusion, a coarse-to-fine pipeline that marries the denoising diffusion model with UV mapping to generate consistent and texture images in space. start introducing point synthesize low-frequency components our tailored style guidance tackle biased color distribution. The derived coarse offers global consistency serves as condition for subsequent stage, aiding regularizing image. Then, hybrid...
The development of a solid-state electrolyte (SSE) is crucial for overcoming the side reactions metal potassium anodes and advancing progress K-ion batteries (KIBs). Exploring diffusion mechanism K ion in SSE important deepening our understanding promoting its development. In this study, we conducted static calculations utilized deep potential molecular dynamics (DeepMD) to investigate behavior cubic K3SbS4. original K3SbS4 exhibited poor ionic conductivity, but discovered that introducing...
Long-range temporal alignment is critical yet challenging for video restoration tasks. Recently, some works attempt to divide the long-range into several sub-alignments and handle them progressively. Although this operation helpful in modeling distant correspondences, error accumulation inevitable due propagation mechanism. In work, we present a novel, generic iterative module which employs gradual refinement scheme sub-alignments, yielding more accurate motion compensation. To further...
Neural radiance fields (NeRF) show great success in novel view synthesis. However, real-world scenes, recovering high-quality details from the source images is still challenging for existing NeRF-based approaches, due to potential imperfect calibration information and scene representation inaccuracy. Even with training frames, synthetic views produced by NeRF models suffer notable rendering artifacts, such as noise, blur, etc. Towards improve synthesis quality of we propose NeRFLiX, a...
Abstract The von Neumann bottleneck has long been a significant obstacle to the advancement of era intelligent computing. Neuromorphic devices are considered promising solution overcome this bottleneck. These draw inspiration from information processing and computing capabilities neurons in human brain. Nevertheless, biomimetic synaptic used neural network encounter challenges, including high nonlinearity regulation, limited abundance state conductance, restrictions unidirectional...
For video frame interpolation (VFI), existing deep-learning-based approaches strongly rely on the ground-truth (GT) intermediate frames, which sometimes ignore non-unique nature of motion judging from given adjacent frames. As a result, these methods tend to produce averaged solutions that are not clear enough. To alleviate this issue, we propose relax requirement reconstructing an as close GT possible. Towards end, develop texture consistency loss (TCL) upon assumption interpolated content...
A series of four 1,3,4-oxadiazole-based thermally activated delayed fluorescence (TADF) derivatives are reported as emitters for organic light emitting diodes (OLEDs). As a function the nature substituent on weak 1,3,4-oxadiazole acceptor, their emission color could be tuned from sky-blue to blue. The highly twisted conformation between carbazoles and oxadiazoles results in effective separation highest occupied lowest unoccupied molecular orbitals resulting small singlet–triplet splitting....