A5068251664- Data Management and Algorithms
- Blind Source Separation Techniques
- Human Pose and Action Recognition
- Advanced Image and Video Retrieval Techniques
- EEG and Brain-Computer Interfaces
- Time Series Analysis and Forecasting
- Advanced Data Compression Techniques
- Advanced Vision and Imaging
- Video Surveillance and Tracking Methods
- ECG Monitoring and Analysis
- Data Mining Algorithms and Applications
- Radiomics and Machine Learning in Medical Imaging
- Image and Signal Denoising Methods
- Distributed systems and fault tolerance
- Cloud Computing and Resource Management
- Parallel Computing and Optimization Techniques
- Algorithms and Data Compression
- Image Retrieval and Classification Techniques
- Image Enhancement Techniques
- Distributed and Parallel Computing Systems
- Gait Recognition and Analysis
- COVID-19 diagnosis using AI
- Privacy-Preserving Technologies in Data
- Data Quality and Management
- Hate Speech and Cyberbullying Detection
Ningbo University of Technology
2016-2024
Zhejiang University
2013-2023
Hebei Science and Technology Department
2023
Hebei Academy of Sciences
2014-2023
Australian e-Health Research Centre
2005-2022
Commonwealth Scientific and Industrial Research Organisation
2006-2022
Hebei University of Economics and Business
2022
Griffith University
2021-2022
Ningbo University
2019-2021
Ningbo Institute of Industrial Technology
2021
Online human gesture recognition has a wide range of applications in computer vision, especially human-computer interaction applications. Recent introduction cost-effective depth cameras brings on new trend research body-movement recognition. However, there are two major challenges: i) how to continuously recognize gestures from unsegmented streams, and ii) differentiate different styles same other types gestures. In this paper, we solve these problems with effective efficient feature...
This paper presents a semi-supervised method for categorizing human actions using multiple visual features. The proposed algorithm simultaneously learns features from small number of labeled videos, and automatically utilizes data distributions between unlabeled to boost the recognition performance. Shared structural analysis is applied in our approach discover common subspace shared by each type feature. In subspace, able characterize more discriminative information feature type....
<title>Abstract</title> Objective To propose a deep learning model and explore its performance in the auxiliary diagnosis of lung cancer associated with cystic airspaces (LCCA) computed tomography (CT) images. Methods This study is retrospective analysis that incorporated total 342 CT images, comprising 272 images from patients diagnosed LCCA 70 pulmonary bulla. A named LungSSFNet, developed based on nnUnet, was utilized for image recognition segmentation by experienced thoracic surgeons....
The scarcity data of medical field brings the collaborative training in vision-language pre-training (VLP) cross different clients. Therefore, VLP faces two challenges: First, requires privacy, thus can not directly shared across Second, distribution institutes is typically heterogeneous, hindering local model alignment and representation capabilities. To simultaneously overcome these challenges, we propose framework called personalized selector with fused multimodal information (PMS-FM)....
Human action recognition in videos draws strong research interest computer vision because of its promising applications for video surveillance, annotation, interactive gaming, etc. However, the amount data containing human actions is increasing exponentially, which makes management these resources a challenging task. Given database with huge volumes unlabeled videos, it prohibitive to manually assign specific types videos. Considering that much easier obtain small number labeled practical...
The surge of medical and e-commerce applications has generated tremendous amount data, which brings people to a so-called "Big Data" era. Different from traditional large data sets, the term not only means size volume but also indicates high velocity generation. However, current mining analytical techniques are facing challenge dealing with in short period time. This paper explores efficiency utilizing Normal Distribution (ND) method for splitting processing cloud environment, can provide...
Delineating the crucial waves in electrocardiogram records is a paramount work for automatic diagnosis system of heart diseases. In this paper, novel method described to determine boundaries and peaks P waves, QRS complexes T by utilizing twelve-lead signals. It avoids difficulty setting thresholds when determining also trouble selection wavelet basis as wavelet-based does. The signals are first preprocessed bandpass filter. After that, locations identified. And based on locations, adaptive...
Online human gesture recognition has a wide range of applications in computer vision, especially human-computer interaction applications. The recent introduction cost-effective depth cameras brings new trend research on body-movement recognition. However, there are two major challenges: (i) how to continuously detect gestures from unsegmented streams, and (ii) differentiate different styles the same other types gestures. In this article, we solve these problems with effective efficient...
This paper addresses the challenges in detecting potential correlation between numerical data streams, which facilitates research of stream mining and pattern discovery. We focus on local with delay, may occur burst at different time last for a limited period. The uncertainty occurrence delay make it difficult to monitor online. Furthermore, conventional measure lacks ability reflecting visual linearity, is more desirable reality. proposes effective methods continuously detect streams. Our...
Given a database, the view maintenance problem is concerned with efficient computation of new contents given when updates to database happen. We consider for situation contains weighted graph and either transitive closure or answer all-pairs shortest-distance ( APSD ). give incremental algorithms , which support both edge insertions deletions. For closure, algorithm applicable more general class graphs than those previously explored. Our use first-order queries, along addition (+) less-than...
State-of-the-art studies on cyberbullying detection, using text classification, predominantly take it for granted that streaming can be completely labelled. However, the rapid growth of unlabelled data generated in real time from online content