A5101520408- AI in cancer detection
- Radiomics and Machine Learning in Medical Imaging
- Advanced Image Processing Techniques
- Domain Adaptation and Few-Shot Learning
- Advanced Neural Network Applications
- COVID-19 diagnosis using AI
- Advanced Image and Video Retrieval Techniques
- Advanced Image Fusion Techniques
- Image Enhancement Techniques
- Generative Adversarial Networks and Image Synthesis
- Advanced Vision and Imaging
- Complex Network Analysis Techniques
- Multimodal Machine Learning Applications
- Medical Imaging and Analysis
- Service-Oriented Architecture and Web Services
- Video Coding and Compression Technologies
- Surface Modification and Superhydrophobicity
- Image and Video Quality Assessment
- Advanced Graph Neural Networks
- Recommender Systems and Techniques
- Industrial Vision Systems and Defect Detection
- Remote-Sensing Image Classification
- Web Data Mining and Analysis
- Imbalanced Data Classification Techniques
- Dental Radiography and Imaging
Tianjin University
2018-2024
Jilin University
2023-2024
Shenzhen University
2023
Peking University
2015-2021
Xinjiang University
2021
Chongqing University
2021
Guilin University of Electronic Technology
2020
Ocean University of China
2019
Nankai University
2018
Zhengzhou University of Aeronautics
2010-2012
Implant-associated infections and excessive immune responses are two major postsurgical issues for successful implantation. However, conventional strategies including antibiotic treatment inflammatory regulation always compromised due to the comodification of various biochemical agents instances functional interference. It is imperative provide implant surfaces with satisfactory antibacterial anti-inflammatory properties. Here, a dual-effect nanostructured polyetheretherketone (PEEK) surface...
Ground object detection, based on remote sensing satellite imagery, provides the groundwork of numerous applications, so detection accuracy is vital importance. The background images complex, size various, and there are many small objects. In view above problems, a multi-attention method (MA-FPN) multi-scale proposed in this paper, which can effectively make network pay attention to location reduce loss information. According feature pyramid (FPN), we firstly put forward global spatial...
It has proposed an adaptive threshold edge detection algorithm in this paper, which applies the bilateral filtering that advantages of edge-preserving and noise-removing firstly. Then it uses OTSU, is based on gradient magnitude to maximize separability resultant classes, determine low high thresholds canny operator. Finally, connection are performed. The experimental results show practical reliable.
Now, the security and privacy-preserving of Internet Things are receiving more attention. This article proposes a new method privacy preserving by combining differentiated nodes, precision clustering, RSA, multi-signature, blockchain. To calculate node rank value in Things, number nodes' links is taken as weight nodes based on PageRank. increases values' difference between nodes. Accurate clustering determines initial central k-means values, which effectively differentiates active inactive...
Accurate and real-time detection of airplanes, cars, ships in remote sensing images is an important but challenging task that plays role both military civilian life. The most issues posed by this are the intensive tiny size objects complexity application scenarios. In letter, we propose a multi-vision small object detector can rapidly accurately detect images. We make following three contributions: multiscale residual block (MRB) proposed, whereby dilated convolution employed cascade to...
Endoscopic images captured under low-light enclosed intestinal environment usually have poor visibility (manifested as uneven illumination and noise), affecting the work efficiency of physicians accuracy lesion detection. To improve image quality, literature has reported many enhancement (LIE) methods. However, most methods do not perform well in handling endoscopic (LEIE) task, bringing additional artifacts or amplifying noise. In this paper, we propose a novel deep pyramid network (DPENet)...
The generation-based data augmentation method can overcome the challenge caused by imbalance of medical image to a certain extent. However, most current research focus on images with unified structure which are easy learn. What is different that ultrasound structurally inadequate, making it difficult for be captured generative network, resulting in generated lacks structural legitimacy. Therefore, Progressive Generative Adversarial Method Structurally Inadequate Medical Image Data...
The key to a recommendation system is the prediction of users' preferences. Personalized for many online music applications depends on both long-term as well short-term In this paper, we propose novel personalized next-song that jointly consider and preferences in its design. To depict preferences, divide user network into communities according their similar activities; meanwhile fuse such with Markov chain make sequential recommendations are better tailored well. We perform an experimental...
In this paper, a novel QP variable convolutional neural network based in-loop filter is proposed for VVC intra coding. To avoid training and deploying multiple networks, we develop an efficient attention module (QPAM) which can capture compression noise levels different QPs emphasize meaningful features along channel dimension. Then embed QPAM into the residual block, on it, design architecture that equipped with controllability QPs. make model focus more examples have artifacts or hard to...
The class imbalance problem, which is prevalent in medical image datasets, seriously affects the diagnostic effectiveness of deep learning-based network models. Recently, method based on two-stage learning has produced promising results solving imbalance. In learning, unbiased classifiers been well studied, but representation imbalanced data still being explored. this paper, we focus stage class-imbalanced and propose a novel balanced discriminative contrastive (BDCL) method. Compared with...
Recognizing text in natural scenes is still a very challenging task, due to arbitrary shapes, varying fonts, complex backgrounds and so on. Recently, some recognizers utilize Spatial Transform Network (STN) rectify irregular instances achieve promising results. However, their robustness accuracy are limited, since rectification performance can be easily degraded by samples. To tackle this issue, we propose simple yet effective two-dimensional (2D) character attention module, which enhance...
With the maturity of artificial intelligence, AI-aided diagnosis technology is gradually widely applied in clinical medicine. However, for same pathological tissue, medical images produced by different types instruments usually possess data distributions. Because domain shift phenomenon, cannot accurately diagnose other domains, which a waste precious images. This paper proposes Multi-Scale Self-Attention Unsupervised Domain Adaptive framework (MSDAN), consists three modules. First,...
Semi-supervised domain generalization (SSDG) aims to build a domain-generalized model using partially labeled data from source domains. Mainstream SSDG methods follow the augmentation consistency in FixMatch. However, extraction of domain-invariant features may be challenging due absence feature-based operations, further leading overfitting classifier. To this end, we propose Multi-level Augmentation Consistency Learning (MACMatch), which improves feature extractor and classifier through...
Deep learning-based segmentation methods are widely utilized for detecting lesions in ultrasound images. Throughout the imaging procedure, attenuation and scattering of waves cause contour blurring formation artifacts, limiting clarity acquired To overcome this challenge, we propose a contour-based probabilistic model CP-UNet, which guides network to enhance its focus on during decoding. We design novel down-sampling module enable probability distribution modeling encoding stages acquire...