A5060090179- Advanced Image and Video Retrieval Techniques
- Advanced MIMO Systems Optimization
- Cooperative Communication and Network Coding
- Visual Attention and Saliency Detection
- Advanced Neural Network Applications
- Advanced Wireless Network Optimization
- Image and Signal Denoising Methods
- Medical Image Segmentation Techniques
- Optical Coherence Tomography Applications
- Image Processing Techniques and Applications
- Olfactory and Sensory Function Studies
- Image Enhancement Techniques
- Image and Video Quality Assessment
- Advanced Image Processing Techniques
- Image Retrieval and Classification Techniques
- Face Recognition and Perception
- Advanced Photonic Communication Systems
- Advanced Wireless Communication Technologies
- Ocular and Laser Science Research
- Glaucoma and retinal disorders
- Medical Imaging and Analysis
- AI in cancer detection
- Laser Material Processing Techniques
- Ultrasound Imaging and Elastography
- Advanced Wireless Communication Techniques
Wuhan University of Technology
2024-2025
University of California, Irvine
2025
China Academy of Space Technology
2021-2024
Huazhong University of Science and Technology
2012-2024
Nanjing University of Posts and Telecommunications
2014-2024
Chinese Academy of Surveying and Mapping
2024
Northwestern Polytechnical University
2023
Wuzhou University
2023
Zero to Three
2022
Anhui University of Technology
2020-2021
In daily life, there are a variety of complex sound sources. It is important to effectively detect certain sounds in some situations. With the outbreak COVID-19, it necessary distinguish coughing, estimate suspected patients population. this paper, we propose method for cough recognition based on Mel-spectrogram and Convolutional Neural Network called Cough Recognition (CRN), which can sounds.
In the field of disease diagnosis where only a small dataset medical images may be accessible, light-weight convolutional neural network (CNN) has become popular because it can help to avoid over-fitting problem and improve computational efficiency. However, feature extraction capability CNN is inferior that heavy-weight counterpart. Although attention mechanism provides feasible solution this problem, existing modules, such as squeeze excitation module block module, have insufficient...
Onboard real-time object detection in remote sensing images is a crucial but challenging task this computation-constrained scenario. This not only requires the algorithm to yield excellent performance also requests limited time and space complexity of algorithm. However, previous convolutional neural networks (CNN) based detectors for suffer from heavy computational cost, which hinders them being deployed on satellites. Moreover, an onboard detector desired detect objects at vastly different...
The predicted capacity gain of a traditional co- located MIMO system is often severely limited in realistic propagation scenarios, especially when the number antennas becomes large. Recently, generalized paradigm for multiple-antenna communications, distributed MIMO, proposed as remedy. In this paper, through asymptotic large-system analysis, we provide solid justifications on advantages over co-located communication channels are subject to spatial correlation and shadow fading. We also...
This paper demonstrates the impact of 3D effects on performance parameters in small-sized Time Delay Integration (TDI) image sensor pixels. In this paper, 2D and simulation models 3.5 μm × TDI pixels were constructed, utilizing a three-phase pixel structure integrated with lateral anti-blooming structure. The experiments reveal limitations traditional by comparing results. research validates influence barrier height full well potential proposes methods to optimize operating voltage To verify...
Three-dimensional (3D) medical image segmentation typically demands extensive labeled training samples, which is prohibitively time-consuming and requires significant expertise. Although this demand can be mitigated by special learning paradigms such as semi-supervised learning, the cost still high due to reader-unfriendly 3D data structure. In paper, we seek a solution of robust using extremely simplified annotation that delineates only single slice per each volume for subset samples. To...
Through the analysis of probability density function squared largest singular value a complex Gaussian matrix at origin and tail, we obtain two asymptotic results related to multi-input multi-output (MIMO) maximum-ratio-transmission/maximum-ratio-combining (MRT/MRC) systems. One is error performance (in terms SNR) in single-user system, other system capacity number users) multiuser scenario when diversity exploited. Similar are also obtained for MIMO schemes, space-time block coding...
图像模糊是指在图像捕捉或传输过程中,由于镜头或相机运动、光照条件等因素导致图像失去清晰度和细节,从而影响图像的质量和可用性。为了消除这种影响,图像去模糊技术应运而生。其目的在于通过构建计算机数学模型来衡量图像的模糊信息,从而自动预测去模糊后的清晰图像。图像去模糊算法的研究发展不仅为计算机视觉领域的其他任务提供了便利,同时也为生活领域提供了便捷和保障,如安全监控等。1)回顾了整个图像去模糊领域的发展历程,对盲图像去模糊和非盲图像去模糊中具有影响力的算法进行论述和分析。2)讨论了图像模糊的常见原因以及去模糊图像的质量评价方法。3)全面阐述了传统方法和基于深度学习方法的基本思想,并针对图像非盲去模糊和图像盲去模糊两方面的一些文献进行了综述。其中,基于深度学习的方法包括基于卷积神经网络、基于循环神经网络、基于生成式对抗网络和基于Transformer的方法等。4)简要介绍了图像去模糊领域的常用数据集并比较分析了一些代表性图像去模糊算法的性能。5)探讨了图像去模糊领域所面临的挑战,并对未来的研究方法进行了展望。;Image blurring refers to the loss of...
Optical coherence tomography (OCT) has found wide application to the diagnosis of ophthalmic diseases, but quality OCT images is degraded by speckle noise. The convolutional neural network (CNN) based methods have attracted much attention in image despeckling. However, these generally need noisy-clean pairs for training and they are difficult capture global context information effectively. To address issues, we proposed a novel unsupervised despeckling method. This method uses cross-scale...
This Letter presents a guided filtering (GF)-based nonlocal means (NLM) method for despeckling of optical coherence tomography (OCT) images. Unlike existing NLM methods that determine weights using image intensities or features, the proposed first uses GF to capture both grayscale information and features input then introduces them into accurate weight computation. The boosting iterative strategies are further incorporated ensure performance. Experiments on real OCT images demonstrate our...
With the rapid development of wearable technologies, devices are gradually entering people's daily life. As a novel way human-computer interaction, have brought more and convenience assistance to people than ever before. In this paper, we firstly make classification based on their functions how they worn. An introduction current is also presented. Then, challenges in area possible solutions discussed. Finally, conclusions drawn future trends technologies.
Although convolutional neural networks (CNNs) have made significant progress, their deployment onboard is still challenging because of complexity and high processing cost. Tensors provide a natural compact representation CNN weights via suitable low-rank approximations. A novel decomposed module called DecomResnet based on Tucker decomposition was proposed to deploy object detection model satellite. We remote sensing image compression framework which consisted four steps, namely (1)...
Visual objects are recognized by their features. Whereas some features based on simple components (i.e., local features, such as orientation of line segments), the whole object global an having a hole in it). Over past five decades, behavioral, physiological, anatomical and computational studies have established general model vision, which starts from extracting lower visual pathways followed feature integration process that extracts higher pathways. This local-to-global is successful...
An object-based attention model to predict visual saliency using contrast against the ‘background prototypes’ is presented. The proposed automatically identifies a series of regions far away from image centre as background prototypes. then calculated colour these Promising experimental results demonstrate effectiveness in terms detection accuracy and implementation efficiency.