A5007736553- Context-Aware Activity Recognition Systems
- IoT and Edge/Fog Computing
- Non-Invasive Vital Sign Monitoring
- Anomaly Detection Techniques and Applications
- Human Pose and Action Recognition
- Medical Imaging Techniques and Applications
- Advanced MRI Techniques and Applications
- Higher Education and Teaching Methods
- Geothermal Energy Systems and Applications
- Heat Transfer and Boiling Studies
- Reconstructive Facial Surgery Techniques
- Respiratory viral infections research
- Radiomics and Machine Learning in Medical Imaging
- Facial Rejuvenation and Surgery Techniques
- Probabilistic and Robust Engineering Design
- Digital Imaging for Blood Diseases
- Retinal Imaging and Analysis
- CO2 Sequestration and Geologic Interactions
- Viral gastroenteritis research and epidemiology
- Industrial Vision Systems and Defect Detection
- Fatigue and fracture mechanics
- Gait Recognition and Analysis
- Cleft Lip and Palate Research
- Solar Thermal and Photovoltaic Systems
- Uterine Myomas and Treatments
Southeast University
2023-2024
University of Macau
2021-2024
Nanjing Normal University
2021-2024
China Power Engineering Consulting Group (China)
2024
Jilin University
2016-2024
Guangzhou Institute of Energy Conversion
2018-2024
Chinese Academy of Sciences
2014-2024
National University of Defense Technology
2024
Shaanxi Polytechnic Institute
2024
Tsinghua University
2023
Recently, human activity recognition (HAR) that uses wearable sensors has become a research hotspot because its wide applications in real-world scenarios. Essentially, HAR can be treated as multi-channel time series classification problem, where different channels may come from heterogeneous sensor modalities. Deep learning, especially convolutional neural networks (CNNs) have made breakthroughs ubiquitous scenario. Various normalization methods enable layers of to learn more independently...
Due to rapid development of sensor technology, human activity recognition (HAR) using wearable inertial sensors has recently become a new research hotspot. Deep learning, especially convolutional neural network (CNN) that can automatically learn intricate features have gained lot attention in ubiquitous HAR task. Most existing CNNs process input by extracting channel-wise features, and the information from each channel be separately propagated hierarchical way lower layers higher layers. As...
Recently, the state-of-the-art performance in various sensor-based human activity recognition (HAR) tasks has been acquired by deep learning, which can extract automatically features from raw data. In standard convolutional neural networks (CNNs), there is usually same receptive field (RF) size of artificial neurons within each feature layer. It well known that RF able to change adaptively according stimulus, rarely exploited HAR. this article, a new multibranch CNN introduced, utilizes...
Recent years have witnessed significant success of convolutional neural networks (CNNs) in human activity recognition (HAR) using wearable sensors. Nevertheless, prior works an obvious drawback. An sample may contain heterogeneous sensor modalities from different body parts. Moreover, the significance each modality will change over time. Because a normal convolution filter usually samples data at fixed regular grid, it is hard to capture salient features activities along or time intervals....
During the past decade, human activity recognition ( HAR ) using wearable sensors has become a new research hot spot due to its extensive use in various application domains such as healthcare, fitness, smart homes, and eldercare. Deep neural networks, especially convolutional networks CNNs ), have gained lot of attention scenario. Despite exceptional performance, with heavy overhead is not best option for task limitation computing resource on embedded devices. As far we know, there are many...
Recently, the state-of-the-art performance in various sensor based human activity recognition (HAR) tasks have been acquired by deep learning, which can extract automatically features from raw data. In order to obtain best accuracy, many static layers always used train neural networks, and their weight connectivity network remains unchanged. Pursuing accuracy mobile platforms with a very limited computational budget at millions of FLOPs is impractical. this paper, we make use shallow...
Human Activity Recognition (HAR) aims to recognize activities by training models on massive sensor data. In real-world deployment, a crucial aspect of HAR that has been largely overlooked is the test sets may have different distributions from due inter-subject variability including age, gender, behavioral habits, etc., which leads poor generalization performance. One promising solution learn domain-invariant representations enable model generalize an unseen distribution. However, most...
In few-shot action recognition (FSAR), long sub-sequences of video naturally express entire actions more effectively. However, the high computational complexity mainstream Transformer-based methods limits their application. Recent Mamba demonstrates efficiency in modeling sequences, but directly applying to FSAR overlooks importance local feature and alignment. Moreover, within same class accumulate intra-class variance, which adversely impacts performance. To solve these challenges, we...
To date, convolutional neural networks have played a dominant role in sensor-based human activity recognition (HAR) scenarios. In 2021, researchers from four institutions almost simultaneously released their newest work to arXiv.org, where each of them independently presented new network architectures mainly consisting linear layers. This arouses heated debate whether the current research hotspot deep learning is returning MLPs. Inspired by recent success achieved MLPs, this paper, we first...
A generalized stress-strength interference (SSI) reliability model to consider stochastic loading and strength aging degradation is presented in this paper. This conforms previous models for special cases, but also demonstrates the weakness of those when multiple elements exist. It can be used any nonhomogeneous Poisson process, kind model. To solve SSI equation, a numerical recurrence formula based on Gauss-Legendre quadrature calculate integrations random variable vector. Numerical...
Activity recognition plays a critical role in various applications, such as medical monitoring and rehabilitation. Deep learning has recently made great development the wearable-based human activity (HAR) area. However, real HAR applications should be adaptive flexible to available computational budget. So far, this problem rarely been explored. In contrast existing deep studies focusing on static networks, article aims investigate which can adjust their structure conditioned computing...
Clitoria ternatea (CT) flowers are rich in phytochemicals. An innovative approach was taken to utilize CT flower extract (CTFE) as a functional ingredient with natural pigment by incorporating it into noodles. The aim of this study examine the effect CTFE amount (0-30%) on color, texture, phytochemicals, and sensory quality both dried cooked Dried noodles 30% had highest total anthocyanins (9.48 μg/g), polyphenols (612 DPPH radical scavenging capacity (165 μg TE/g), reducing power (2203...