A5052461790- Handwritten Text Recognition Techniques
- Neural dynamics and brain function
- Image Processing and 3D Reconstruction
- Electrochemical Analysis and Applications
- Generative Adversarial Networks and Image Synthesis
- Functional Brain Connectivity Studies
- Video Surveillance and Tracking Methods
- Cryptographic Implementations and Security
- Image and Signal Denoising Methods
- Blind Source Separation Techniques
- Human Motion and Animation
- Geoscience and Mining Technology
- Neural Networks Stability and Synchronization
- Advanced Sensor and Control Systems
- Natural Language Processing Techniques
- Advanced Image Fusion Techniques
- Fractal and DNA sequence analysis
- Non-Destructive Testing Techniques
- Advanced Computational Techniques and Applications
- Chaos control and synchronization
- Quantum chaos and dynamical systems
- Advanced Algorithms and Applications
- Chaos-based Image/Signal Encryption
Peking University
2019-2024
Lanzhou University
2016-2018
Horological Research Institute of Light Industry
2005
Abstract Despite the recent impressive development of deep neural networks, using learning based methods to generate large‐scale Chinese fonts is still a rather challenging task due huge number intricate glyphs, e.g., official standard charset GB18030‐2000 consists 27,533 characters. Until now, most existing models for this adopt Convolutional Neural Networks (CNNs) bitmap images characters CNN models' remarkable success in various applications. However, focus more on image‐level features...
Abstract In this paper, we propose a novel Sequence‐to‐Sequence model based on metric‐based meta learning for the arbitrary style transfer of online Chinese handwritings. Unlike most existing methods that treat handwritings as images and are unable to reflect human writing process, proposed directly handles sequential Generally, our consists three sub‐models: content encoder, encoder decoder, which all Recurrent Neural Networks. order adaptively obtain information, introduce an...
In this paper, the technique of biomedical image is presented by employing cellular neural networks (CNN) and linear matrix inequality (LMI). The main objective to obtain templates. Based on Cellular a high-speed parallel processor, paper proposes optimized templates for edge detection.
It remains unclear whether brain networks are altered during conversion blindness in electroencephalogram (EEG) representation, which is of significance both on improving clinical management blindness, and providing objective evidence for judicial disputes. Functional network was constructed coherence extracted from scalp EEGs, conventional metrics were analyzed various frequency bands. Performance indices using relevant global features evaluated with nonlinear linear classifiers. To...