A5100753233- Remote-Sensing Image Classification
- Remote Sensing and Land Use
- Hand Gesture Recognition Systems
- Face recognition and analysis
- Automated Road and Building Extraction
- Emotion and Mood Recognition
- Geological and Geochemical Analysis
- Advanced Neural Network Applications
- Corneal surgery and disorders
- Remote Sensing and LiDAR Applications
- Ophthalmology and Visual Impairment Studies
- Air Quality and Health Impacts
- Human Pose and Action Recognition
- Advanced Image and Video Retrieval Techniques
- Evaluation Methods in Various Fields
- Flood Risk Assessment and Management
- Face and Expression Recognition
- Human Mobility and Location-Based Analysis
- Video Surveillance and Tracking Methods
- Hydrological Forecasting Using AI
- Structural Health Monitoring Techniques
- Advanced Sensor and Control Systems
- earthquake and tectonic studies
- Gaze Tracking and Assistive Technology
- High-pressure geophysics and materials
Nanyang Technological University
2024-2025
China University of Geosciences
2013-2025
South China University of Technology
2010-2025
Tianjin Medical University
2017-2024
Henan University of Economic and Law
2024
Chinese Center For Disease Control and Prevention
2021-2024
Jiamusi University
2024
Centre National de la Recherche Scientifique
2024
Tohoku University
2024
Tianjin Medical University General Hospital
2023-2024
Building extraction based on high-resolution remote sensing imagery has been widely used in automatic surveying and mapping. However, few methods have developed for building instance extraction, i.e., extracting each building's footprint separately, which is required a number of applications, such as the smallest unit cadastral database. In there are two challenges: 1) buildings with various scales exist 2) precise footprints difficult to extract due blurry boundaries. this article, solve...
Significant advancements in RGB-D semantic segmentation have been made owing to the increasing availability of robust depth information. Most researchers combined with RGB data capture complementary information images. Although this approach improves performance, it requires excessive model parameters. To address problem, we propose DGPINet-KD, a deep-guided and progressive integration network knowledge distillation (KD) for indoor scene analysis. First, used branching attention guidance...
Point cloud denoising is a fundamental and challenging problem in geometry processing. Existing methods typically involve direct of noisy input or filtering raw normals followed by point position updates. Recognizing the crucial relationship between normal filtering, we re-examine this from multitask perspective propose an end-to-end network called PCDNF for joint filtering-based denoising. We introduce auxiliary task to enhance network's ability remove noise while preserving geometric...
Recently, methods based on convolutional neural networks have achieved good results in the dense prediction of remote-sensing images, particularly when employing normalized digital surface models. However, most existing use multiscale convolution and attention to mine multimodal feature information without considering differences complementarities between two features. Moreover, previous studies prioritized model segmentation performance ignored parametric issues, which makes it difficult...
Although tremendous strides have been made in facial expression recognition(FER), recognizing expressions non-frontal views remains an open challenge due to the limited access large scale training data with various poses. To make full use of data, we propose a novel multi-channel pose-aware convolution neural network (MPCNN) that consists three parts: feature extraction, jointly multi-scale fusion, and recognition. The extraction part has 3 sub-CNNs it learns convolutional features from...
Deep learning-based building instance extraction on remote sensing imagery (RSI) has achieved tremendous success under the large-scale labeled training data. However, multi-target domain adaptation (MD-BIE) is still a challenge task that involves transferring knowledge from source to multiple unlabeled target domains, which poses various semantic gaps between and within <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g</i> ., style,...
There is increasing evidence on the link between environmental factors and myopia in children adolescents, yet with inconsistent conclusions. We investigated associations socioeconomic inequalities green space school-aged students participating Tianjin Child Adolescent Research of Eye (TCARE) study.
Environmental contaminants (ECs) are increasingly recognized as crucial drivers of dyslipidemia and cardiovascular disease (CVD), but the comprehensive impact spectrum interlinking mechanisms remain uncertain.
L. monocytogenes is a dangerous food-borne pathogen threatening global food safety, and its antibiotic resistance on the rise. Hop extracts their derivatives were recently found to have antibacterial activity against monocytogenes, while antimicrobial mechanism remains unclear. In this study, new hop derivative, hexahydro-colupulone (HHCL), was prepared based extracts. HHCL inhibited growth of at surprisingly low concentrations (MIC = 0.4 μg/mL). significantly survival in apple juice...
The purpose of this research was to investigate the impact dietary supplementation Caragana korshinskii tannin (CKT) on rumen fermentation, methane emission, methanogen community and metabolome in sheep. A total 15 crossbred sheep Dumont breed with similar body conditions, were divided into three groups ( n = 5), which fed CKT addition at 0, 2 4%/kg DM. study spanned a 74 days, 14-day period dedicated adaptation subsequent 60-day for conducting treatments. results indicated that levels...
Particulate matter (PM) has been found to elevate the risk of pulmonary embolism (PE) onset. Among contributors PM, dust PM stands as second natural source, and its emissions are escalating due climate change. Despite this, information on effect PE onset is scarce. Hence, this study aims investigate impact PM10, PM2.5–10, PM2.5 A nationwide time-stratified case-crossover was conducted between 2015 2020, using data from 18,616 cases across 1921 hospitals. The analysis employed a conditional...
The semantic segmentation of remote sensing images is crucial for computer perception tasks. Integrating dual-modal information enhances understanding. However, existing methods often suffer from incomplete feature (features without integrity), leading to inadequate pixels near object boundaries. This study introduces the concept integrity in and presents a complete learning network using contextual semantics multiscale decoding process. Specifically, we propose frequency-aware (FILNet) that...
In recent years, significant progress has been made in arbitrary motion style transfer. However, many existing methods primarily focus on end-to-end processes and overlook the differences similarities among generated motions, thus failing to fully leverage information. To address this limitation, we employ contrastive learning bring motions closer input further away from others. This enables model construct better representations optimize quality of stylization. Additionally, often struggle...