A5100382233- Human Pose and Action Recognition
- Image and Video Quality Assessment
- Advanced Image Processing Techniques
- Internet Traffic Analysis and Secure E-voting
- Human Motion and Animation
- Image Enhancement Techniques
- Video Surveillance and Tracking Methods
- Visual Attention and Saliency Detection
- Gait Recognition and Analysis
- Radiomics and Machine Learning in Medical Imaging
- EEG and Brain-Computer Interfaces
- Advanced Vision and Imaging
- Software-Defined Networks and 5G
- Network Security and Intrusion Detection
- Technology-Enhanced Education Studies
- Lung Cancer Diagnosis and Treatment
- Advanced Computational Techniques and Applications
- Metamaterials and Metasurfaces Applications
- Image Retrieval and Classification Techniques
- Blind Source Separation Techniques
- Anomaly Detection Techniques and Applications
- Video Coding and Compression Technologies
- Diabetic Foot Ulcer Assessment and Management
- Plasmonic and Surface Plasmon Research
- Gaze Tracking and Assistive Technology
Jiangnan University
2024-2025
Guilin University of Aerospace Technology
2014-2024
Universiti Putra Malaysia
2023-2024
Guilin University of Electronic Technology
2017-2023
Ningbo University of Technology
2023
State Grid Corporation of China (China)
2020-2023
University of Saskatchewan
2023
Saarland University
2022
Netflix (United States)
2015-2022
Peking University
2022
The purpose of this cluster-randomized control field trial was to examine the extent which kindergarten teachers could learn a promising instructional strategy, wherein reading instruction differentiated based upon students' ongoing assessments language and literacy skills documented child characteristic by (CXI) interactions; test efficacy on outcomes students from culturally diverse backgrounds. study involved 14 schools included 23 treatment (
Background Lymphovascular invasion (LVI) status facilitates the selection of optimal therapeutic strategy for breast cancer patients, but in clinical practice LVI is determined pathological specimens after resection. Purpose To explore use dynamic contrast‐enhanced (DCE)‐magnetic resonance imaging (MRI)‐based radiomics preoperative prediction invasive cancer. Study Type Prospective. Population Ninety training cohort patients (22 LVI‐positive and 68 LVI‐negative) 59 validation 37 were...
In order to undertake a series of case studies aimed at investigating systematic literature reviews, we have developed study protocol template. This paper introduces the template and discusses our experiences using resulting protocol. We suggest that prepare for planning can improve rigour software engineering studies.
Simple quality metrics such as PSNR are known to not correlate well with subjective when tested across a wide spectrum of video content or regime. Recently, efforts have been made in designing objective trained on data, demonstrating better correlation perceived by human. Clearly, the accuracy metric heavily depends data that it is on. In this paper, we propose new approach recover scores from noisy raw measurements, jointly estimating impaired videos, bias and consistency test subjects,...
We propose a new meta-heuristic algorithm named Levy flights multi-verse optimizer (LFMVO), which incorporates into (MVO) to solve numerical and engineering optimization problems. The Original MVO easily falls stagnation when wormholes stochastically re-span number of universes (solutions) around the best universe achieved over course iterations. Since are superior in exploring unknown, large-scale search space, they integrated previous force out stagnation. test this method on three sets 23...
The premise of training an accurate 3D human pose estimation network is the possession huge amount richly annotated data. Nonetheless, manually obtaining rich and annotations is, even not impossible, tedious slow. In this paper, we propose to exploit monocular videos complement dataset for single-image tasks. At beginning, a baseline model trained with small set annotations. By fixing some reliable estimations produced by resulting model, our method automatically collects across entire video...
The paper explores the performance of LLMs in context multi-dimensional analytic writing assessments, i.e. their ability to provide both scores and comments based on multiple assessment criteria. Using a corpus literature reviews written by L2 graduate students assessed human experts against 9 criteria, we prompt several popular perform same task under various conditions. To evaluate quality feedback comments, apply novel comment evaluation framework. This framework is interpretable,...
ABSTRACT Aim To explore the motivations, capacity preparations and career development plans of specialist nurses (SNs) to provide insights for promoting further specialisation in nursing enhancing quality services. Design A descriptive phenomenological qualitative study was conducted. Methods The conducted from April May 2024 at a tertiary hospital China. purposive sampling method used recruit 35 various departments who had completed training obtained necessary qualifications. These SNs...
Many schools are beginning to implement multi-tier response intervention (RTI) models for the prevention of reading difficulties and assist in identification students with learning disabilities (LD). The present study was part our larger ongoing longitudinal RTI investigation within Florida Learning Disabilities Center grant. This used a correlational design, conducted 7 ethnically socio-economically diverse schools. We observed instruction 20 classrooms, examined rates kindergarten Tier 1...
The Netflix ingest and encoding pipeline is a cloud-based platform that generates video encodes for the streaming service. Due to large throughput of system, automated quality assessment source videos generated essential in ensuring experience viewers. This paper discusses motivations integrating production pipeline, outlines currently deployed solutions presents technical challenges improving system.
The common spatial pattern (CSP) is a very effective feature extraction method in motor imagery based brain computer interface (BCI), but its performance depends on the selection of optimal frequency band. Although lot research works have been proposed to improve CSP, most these problems large computation costs and long time. To this end, three new methods CSP non-convex log regularization are paper. Firstly, EEG signals spatially filtered by then proposed. We called them CSP-wavelet,...
This paper aims to analyze the predictive effect of artificial intelligence on user demand in big data social media and provide suggestions for developing enterprise innovation frameworks implementing marketing strategies. In response inconsistency between supply products services market demand, deep learning algorithms have been introduced using analysis. algorithm has improved construct a prediction model based bidirectional long short-term memory (BiLSTM) fused with Word2Vec. The uses...
This study aimed to find suitable source domain data in cross-domain transfer learning extract robust image features. Then, a model was built preoperatively distinguish lung granulomatous nodules (LGNs) from adenocarcinoma (LAC) solitary pulmonary solid (SPSNs).
User expectations have a crucial impact on the levels of quality experience (QoE) that they consider acceptable or satisfying. Measuring acceptability and annoyance has mainly been performed in separate multi-step experiments without any control over participants' expectations. This paper introduces simple methodology to obtain information about both entities single step compares several data processing strategies useful for results interpretation. A specifically designed subjective...
3D human motion capture from monocular RGB images respecting interactions of a subject with complex and possibly deformable environments is very challenging, illposed under-explored problem. Existing methods address it only weakly do not model possible surface deformations often occurring when humans interact scene surfaces. In contrast, this paper proposes MoCapDe-form, i.e., new framework for that the first to explicitly non-rigid improved pose estimation environment reconstruction....
Feature extraction and selection are important parts of motor imagery electroencephalogram (EEG) decoding have always been the focus difficulty brain-computer interface (BCI) system research. In order to improve accuracy EEG reduce model training time, new feature methods proposed in this paper. First, a spatial-frequency method is proposed. The original signal preprocessed, then common spatial pattern (CSP) used for filtering dimensionality reduction. Finally, filter bank decompose...
When developing data-driven video quality assessment algorithms, the size of available ground truth subjective data may hamper generalization capabilities trained models. Nevertheless, if application context is known a priori, leveraging approaches for prediction can deliver promising results. Towards achieving highperforming compression and scaling artifacts, Netflix developed Video Multi-method Assessment Fusion (VMAF) Framework, full-reference system which uses regression scheme to...