A5103326247- Generative Adversarial Networks and Image Synthesis
- Computer Graphics and Visualization Techniques
- Advanced Vision and Imaging
- Retinal Imaging and Analysis
- Digital Media Forensic Detection
- Advanced Steganography and Watermarking Techniques
- Retinal Diseases and Treatments
- Advanced Image Fusion Techniques
- Glaucoma and retinal disorders
- Advanced Image Processing Techniques
- Infrared Target Detection Methodologies
- Advanced Measurement and Detection Methods
- Biometric Identification and Security
- Face recognition and analysis
- Semantic Web and Ontologies
- Visual Attention and Saliency Detection
- Color perception and design
- Image Enhancement Techniques
- Impact of Light on Environment and Health
- Advanced Neural Network Applications
- Neural Networks and Applications
- Privacy-Preserving Technologies in Data
- Stochastic Gradient Optimization Techniques
- Ocular and Laser Science Research
- Image Processing Techniques and Applications
Beihang University
2021-2025
Shandong Provincial Hospital
2025
Shandong University
2025
Tsinghua University
2022
South China Normal University
2022
Automatic generation of fonts can greatly facilitate the font design process, and provide prototypes where designers draw inspiration from. Existing methods are mainly built upon rasterized glyph images to utilize successful convolutional architecture, but ignore vector nature shapes. We present an implicit representation, modeling each as shape primitives enclosed by several quadratic curves. This structured representation is shown be better suited for modeling, enables rendering at...
Pan-sharpening aims to obtain a multi-spectral image of high resolution from inputs spatial panchromatic and low image. In recent years, pan-sharpening methods based on supervised learning have achieved superior performance over traditional methods. However, all these rest upon the assumption that model trained coarse scale can generalize well finer one, which is not always case. To address this problem, we propose novel dual-output cross-scale strategy DOCSNet for pan-sharpening. consists...
Selecting views is one of the most common but overlooked procedures in topics related to 3D scenes. Typically, existing applications and researchers manually select through a trial-and-error process or "preset" direction, such as top-down views. For example, literature for scene synthesis requires visualizing Research on panorama VR also require initial placements cameras, etc. This article presents SceneViewer, an integrated system automatic view selections. Our achieved by applying rules...
Abstract 360° videos enable viewers to watch freely from different directions but inevitably prevent them perceiving all the helpful information. To mitigate this problem, picture-in-picture (PIP) guidance was proposed using preview windows show regions of interest (ROIs) outside current view range. We identify several drawbacks representation and propose a new method for film watching called AdaPIP. AdaPIP enhances traditional PIP by adaptively arranging with changeable ranges sizes. In...
Reward design is a fundamental, yet challenging aspect of reinforcement learning (RL). Researchers typically utilize feedback signals from the environment to handcraft reward function, but this process not always effective due varying scale and intricate dependencies signals. This paper shows by exploiting certain structures, one can ease process. Specifically, we propose hierarchical modeling framework -- HERON for scenarios: (I) The naturally present hierarchy; (II) sparse, with less...
Abstract Quantum communication is a new emerging inter-discipline in which classical theory and quantum mechanics are intertwined. Among them, the utilization of entangled state properties fundamental to realization communication. In this way, we can ensure security confidentiality information during transmission based on scheme, i.e., dense coding scheme. It enables two bits by transmitting only one qubit. past decades, scheme has attracted attention many researchers because its promising...
In this paper, we present a novel implicit glyph shape representation, which models glyphs as primitives enclosed by quadratic curves, and naturally enables generating images at arbitrary high resolutions. Experiments on font reconstruction interpolation tasks verified that structured representation is suitable for describing both structure style features of glyphs. Furthermore, based the proposed design simple yet effective disentangled network challenging one-shot transfer problem, achieve...