A5103235399- Advanced Image Fusion Techniques
- Image Enhancement Techniques
- Image Processing Techniques and Applications
- Video Surveillance and Tracking Methods
- Cell Image Analysis Techniques
- Advanced Vision and Imaging
- Advanced Fluorescence Microscopy Techniques
- Infrared Target Detection Methodologies
- Experimental Learning in Engineering
- Medical Imaging Techniques and Applications
- Advanced Image Processing Techniques
- Fire Detection and Safety Systems
- Human Pose and Action Recognition
- Structural Health Monitoring Techniques
- Infrastructure Maintenance and Monitoring
- Computer Graphics and Visualization Techniques
- Visual Attention and Saliency Detection
- Impact of Light on Environment and Health
- Sensor Technology and Measurement Systems
- Evaluation Methods in Various Fields
- Image and Signal Denoising Methods
- Satellite Image Processing and Photogrammetry
Northeast Forestry University
2024
Zhejiang University
2018-2023
Shenzhen University
2020-2023
Wuhan University
1998-2022
An accurate and fast reconstruction algorithm is crucial for the improvement of temporal resolution in high-density super-resolution microscopy, particularly view challenges associated with live-cell imaging. In this work, we design a deep network based on convolutional neural to take advantage its enhanced ability molecule localization, introduce residual layer into reduce noise. The proposed scheme also incorporates robustness against variations both full width at half maximum (FWHM) pixel...
Compared to color images captured by conventional RGB cameras, monochrome (mono) usually have higher signal-to-noise ratios (SNR) and richer textures due the lack of filter arrays in mono cameras. Therefore, using a mono-color stereo dual-camera system, we can integrate lightness information target with guidance accomplish image enhancement colorization manner. In this work, based on two assumptions, introduce novel probabilistic-concept guided framework. First, adjacent contents similar...
Steel box girder bridges constitute a pivotal structural component in modern bridge engineering, confronting intricate mechanical environments and dynamic conditions during construction, with particularly notable risk of deflection. Risk assessments predominantly rely on traditional analyses empirical judgments, which need help to fully capture the construction changes latent risks. This study introduces an innovative assessment methodology grounded finite element analysis (FEA) optimized by...
In order to efficiently remove honeycomb artifacts and restore images in fiber-bundle-based endomicroscopy, we develop a meta-learning algorithm this work. Two sub-networks are used extract different levels of features. Meta-training is employed train the network with small amount simulated training data, enabling optimal model generalize new tasks not seen set. Numerical results on both USAF target endomicroscopy living mice tissues demonstrate that can high contrast image without pixilated...
In an underwater environment, light always scatters and absorbs as it travels from the object to camera, seriously affecting image quality. As a result, images have poorer contrast, like haze covering them. this paper, modified method for restoration based on dark channel prior (DCP) is presented. Firstly, ambient estimated according difference between blue red channel. Then, attenuations of three RGB channels are obtained separately. Finally, color collection used compensate remaining...
Visual tracking is a visual task that tracks specific target by only giving its first frame location and size. To punish the low-quality but high-scoring results, researchers resorted to foreground reinforcement learning suppress scores of positive samples near edges. However, for training with negative samples, all backgrounds are equally labeled as false. In this way, interdependence difference between background not considered. We interpret underlying reason drifts imbalance embedding...
The soft-argmax operation is widely adopted in neural network-based stereo matching methods to enable differentiable regression of disparity. However, network trained with prone being multimodal due absence explicit constraint the shape probability distribution. Previous leverages Laplacian distribution and cross-entropy for training but failed effectively improve accuracy even compromises efficiency network. In this paper, we conduct a detailed analysis previous distribution-based propose...