A5076764967- Advanced Graph Neural Networks
- Recommender Systems and Techniques
- Asphalt Pavement Performance Evaluation
- Context-Aware Activity Recognition Systems
- Topic Modeling
- Concrete Corrosion and Durability
- Concrete and Cement Materials Research
- Infrastructure Maintenance and Monitoring
- Human Pose and Action Recognition
- Cryospheric studies and observations
- Innovative concrete reinforcement materials
- Evaluation Methods in Various Fields
- Magnesium Oxide Properties and Applications
- Advanced Sensor and Control Systems
- Machine Learning in Healthcare
- Climate change and permafrost
- High-Velocity Impact and Material Behavior
- Structural Response to Dynamic Loads
- Rock Mechanics and Modeling
- Materials Engineering and Processing
- Anomaly Detection Techniques and Applications
- Smart Materials for Construction
- Innovations in Concrete and Construction Materials
- Wireless Sensor Networks and IoT
- Advanced Measurement and Detection Methods
Jiangxi College of Applied Technology
2013-2024
Institute of Computing Technology
2021
Chinese Academy of Sciences
2021
Tianjin University
2020
Xi'an Jiaotong University
2017-2018
Fall detection is an important public healthcare problem. Timely could enable instant delivery of medical service to the injured. A popular nonintrusive solution for fall based on videos obtained through ambient camera, and corresponding methods usually require a large dataset train classifier are inclined be influenced by image quality. However, it hard collect data instead simulated falls recorded construct training dataset, which restricted limited quantity. To address these problems,...
This paper mainly improves the visual geometry group network-16 (VGG-16), which is a classic convolutional neural network (CNN), to classify surface defects on cement concrete bridges in an accurate manner. Specifically, number of fully connected layers was reduced by one, and Softmax classifier replaced with classification layer seven defect tags. The weight parameters pooling were shared pre-trained model, rectified linear unit (ReLU) function taken as activation function. original images...
The deck pavement, as an important structure of the bridge, cushions wheel actions on prevents main girder from rain erosion, and ensures flatness slip resistance for vehicles. asphalt concrete has been widely used in bridge decks. However, conventional crack detection methods cannot identify defects accurately efficiently, due to dark color complexity, different types defects. objective this paper is develop a weakly supervised network segmentation cracks deck. Firstly, data were...
Fall detection is an important problem in the field of public health care, which especially crucial for instant medical service delivery to injured elderly due falls. Ambient camera based fall has been a recognized non-intrusive and publicly acceptable method, where video data employed discriminate event from daily activities. with videos usually requires large dataset extract features train classifier. However, it hard collect free-living environment instead simulated falls by young people...
Learning informative representations (aka. embeddings) of users and items is the core modern recommender systems. Previous works exploit user-item relationships one-hop neighbors in interaction graph to improve quality representation. Recently, research Graph Neural Network (GNN) for recommendation considers implicit collaborative information multi-hop enrich However, most GNN systems do not consider relational which implies expression differences different neighborhood explicitly. The...
Thaw settlement is a major factor influencing deformation, subsidence and collapse, which poses threat to the aircraft safety. Therefore, numerical model of runway structure in permafrost regions established verified by comparing with references. The evolution thaw settlement, transverse non-uniform mechanism are studied based on load A380-800. shows that influence area maximum value settlements increase rapidly first five years slowly over time. mainly concentrated subgrade, gradually...
We present Graph Attention Collaborative Similarity Embedding (GACSE), a new recommendation framework that exploits collaborative information in the user-item bipartite graph for representation learning. Our consists of two parts: first part is to learn explicit filtering such as association through embedding propagation with attention mechanism, and second implicit user-user similarities item-item auxiliary loss. design loss function combines BPR adaptive margin similarity Extensive...
New perspective of the upper structure bridge detection device comprises a body, support, crawling mechanism, gas spring, motor, image acquisition device, buckle plate, chassis and control is installed on cable body bracket, supporting frame set in outside support inner wall creeping mechanism connected with motor side; bracket at bottom chassis, provided photoelectric encoder communication device; part for electrical connection.The utility model can judge external damage cable, which...
Major civil and military infrastructures are usually located in mountainous areas with complex geological structures. When subjected to impact loads caused by operational blasting or seismic activities, the inherent flaws joints hosting rock mass may grow irreversibly eventually cause entire structure collapse. It is, therefore, desirable understand mechanical cracking behaviors of flawed structures under dynamic loads. In this work, cement mortar is used make plate specimens, two different...