A5100668600- Topic Modeling
- Advanced Text Analysis Techniques
- Advanced Vision and Imaging
- Image Processing Techniques and Applications
- Advanced Image Processing Techniques
- Natural Language Processing Techniques
- Image and Signal Denoising Methods
- Video Surveillance and Tracking Methods
- Web Data Mining and Analysis
- Image Enhancement Techniques
- Geological and Geochemical Analysis
- Information Retrieval and Search Behavior
- Geochemistry and Geologic Mapping
- Image Retrieval and Classification Techniques
- earthquake and tectonic studies
- Face recognition and analysis
- Advanced Image Fusion Techniques
- Advanced Image and Video Retrieval Techniques
- Face and Expression Recognition
- Remote-Sensing Image Classification
- Human Pose and Action Recognition
- Sentiment Analysis and Opinion Mining
- Generative Adversarial Networks and Image Synthesis
- Advanced Neural Network Applications
- Robotics and Sensor-Based Localization
Huainan Normal University
2024
Sun Yat-sen University
2022-2024
Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou)
2024
Xi'an Technological University
2020-2023
Nanjing Medical University
2023
Wuyi University
2023
Nanjing University of Posts and Telecommunications
2023
Dongguan University of Technology
2022
University of Toronto
2021
Zhongyuan University of Technology
2020
Abstract The storage of facial images in medical records poses privacy risks due to the sensitive nature personal biometric information that can be extracted from such images. To minimize these risks, we developed a new technology, called digital mask (DM), which is based on three-dimensional reconstruction and deep-learning algorithms irreversibly erase identifiable features, while retaining disease-relevant features needed for diagnosis. In prospective clinical study evaluate technology...
Single image superresolution (SR) aims to construct a high-resolution version from single low-resolution (LR) image. The SR reconstruction is challenging because of the missing details in given LR Thus, it critical explore and exploit effective prior knowledge for boosting performance. In this paper, we propose novel method by exploiting both directional group sparsity gradients features similarity weight estimation. proposed approach based on two observations: 1) most sharp edges are...
The detection of new information in a document stream is an important component many potential applications. In this work, novelty approach based on the identification sentence level patterns proposed. First, information-pattern concept for presented with emphasis general topics (queries) that cannot be simply turned into specific questions whose answers are named entities (NEs). Then we elaborate thorough analysis data from TREC tracks, including lengths, entities, opinion patterns. This...
A hyperspectral image (HSI) contains a great number of spectral bands for each pixel, which will limit the conventional classification methods to distinguish land-cover types pixel. Dimensionality reduction is an effective way improve performance classification. Linear discriminant analysis (LDA) popular dimensionality method HSI classification, assumes all samples obey same distribution. However, different may have contributions in computation scatter matrices. To address problem feature...
Abstract Traditional approaches for monitoring human gait have severe spatial and temporal restrictions with complex analysis methods high cost, which are powerless to promote the development of intelligent life involving fitness, sport training, healthcare. Herein, a portable smart insole system resolution simple manufacturing process measure plantar pressure distribution anytime, anywhere is proposed. An insole‐shaped array 104 piezoresistive sensors highly robust characteristics...
A Hyperspectral Image (HSI) contains a great number of spectral bands for each pixel; however, the spatial resolution HSI is low. image super-resolution effective to enhance while preserving high-spectral-resolution by software techniques. Recently, existing methods have been presented fuse and Multispectral Images (MSI) assuming that MSI same scene required with observed HSI, which limits reconstruction quality. In this paper, new framework based on domain transfer learning proposed...
This study investigates an electromyogram (EMG)-based neural interface toward hand rehabilitation for patients with cerebral palsy (CP). Forty-eight channels of surface EMG signals were recorded from the forearm eight adult subjects CP, while they tried to perform six different grasp patterns. A series myoelectric pattern recognition analyses performed identify movement intention each subject feature sets and classifiers. Our results indicate that across all high accuracies (average overall...
Hyperspectral image (HSI) classification is a widely used application to provide important information of land covers. Each pixel an HSI has hundreds spectral bands, which are often considered as features. However, some features highly correlated and nonlinear. To address these problems, we propose new discrimination analysis framework for based on the Two-stage Subspace Projection (TwoSP) in this paper. First, proposed projects original feature data into higher-dimensional subspace by...
We explore the relationship between time and relevance using TREC ad-hoc queries. A type of query is identified that favors very recent documents. propose a time-based language model approach to retrieval for these show how can be incorporated into both query-likelihood models models. These were used experiments comparing heuristic techniques incorporating document recency in ranking. Our results perform as well or better than best techniques.
In recent years, advancements in high-resolution digital twin platforms or artificial intelligence marine forecasting have led to the increased requirements of oceanic data. However, existing sea surface temperature (SST) products from observations often fail meet researchers’ resolution requirements. Deep learning models serve as practical techniques for improving spatial SST particular, diffusion (DMs) attracted widespread attention due their ability generate more vivid and realistic...
Voltage bias of several hundred volts which are applied between solar cells and module frames may lead to significant power losses, so-called potential-induced degradation (PID), in normal photovoltaic (PV) installations system. Modules minimodules used conduct PID test cells. The procedure is time consuming high cost, cannot be as process monitoring method during fabrication. In this paper, three kinds including minimodule,<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"...
Syntactic information potentially plays a much more important role in question answering than it does retrieval. Although many people have used syntactic evidence Question Answering, there haven't been detailed experiments reported the literature. The aim of experiment described this paper is to study impact particular approach for using on effectiveness. Our results indicate that combination with heuristics ranking potential answers can perform better their own.
Editing faces in videos is a popular yet challenging task computer vision and graphics that encompasses various applications, including facial attractiveness enhancement, makeup transfer, face replacement, expression manipulation. Directly applying the existing warping methods to video editing has major problem of temporal incoherence synthesized videos, which cannot be addressed by simply employing tracking techniques or manual interventions, as it difficult eliminate subtly feature point...
Fuzzy target detection as an important task to reflect the ability of underwater robot, artificial recognition based on image taken by robot has been widely concerned. However, there is no open standard fuzzy data set, and in harsh deep-water environment, it difficult collect large-scale marked optical images. At same time, also hoped that model learn quickly from small samples case few possible. Therefore, combining depth learning transfer learning, a new method improved SSD proposed....
Managing complex enterprise networks requires an understanding at a fine granularity than traditional network monitoring. The ability to correlate and visualize the dynamics inter-relationships among various components such as hosts, users, applications is non-trivial. Network security visualization highlighted topic of research in recent years, existing situation analyzed. paper first proposed awareness model, analysis method, last, designed implemented prototype system based on geographic...