A5100369235- Advanced Neural Network Applications
- Cooperative Communication and Network Coding
- Advanced MIMO Systems Optimization
- Video Surveillance and Tracking Methods
- Advanced Vision and Imaging
- Advanced Image Processing Techniques
- Anomaly Detection Techniques and Applications
- Domain Adaptation and Few-Shot Learning
- Infrared Target Detection Methodologies
- Advanced Image and Video Retrieval Techniques
- Advanced Memory and Neural Computing
- Human Pose and Action Recognition
- Muscle activation and electromyography studies
- Coding theory and cryptography
- Advanced Optical Imaging Technologies
- Augmented Reality Applications
- Advanced Wireless Communication Technologies
- CCD and CMOS Imaging Sensors
- Full-Duplex Wireless Communications
- Advanced Computational Techniques and Applications
- Video Coding and Compression Technologies
- Fire Detection and Safety Systems
- Software System Performance and Reliability
- Cognitive Radio Networks and Spectrum Sensing
- Hand Gesture Recognition Systems
Central China Normal University
2022-2024
Anhui Medical University
2024
Xi'an Jiaotong University
2024
Qingdao Agricultural University
2024
Shanghai Shipbuilding Technology Research Institute
2024
Second Hospital of Anhui Medical University
2024
University of Science and Technology of China
2006-2023
Guangxi University of Finance and Economics
2023
Tencent (China)
2023
Shenzhen Institutes of Advanced Technology
2023
This paper presents a framework for hand gesture recognition based on the information fusion of three-axis accelerometer (ACC) and multichannel electromyography (EMG) sensors. In our framework, start end points meaningful segments are detected automatically by intensity EMG signals. A decision tree multistream hidden Markov models utilized as decision-level to get final results. For sign language (SLR), experimental results classification 72 Chinese Sign Language (CSL) words demonstrate...
A light field is commonly described by a two-plane representation with four dimensions. Refocused 3D contents can be rendered from images. method for capturing these images using cameras microlens arrays. dense sampling of the results in large amounts redundant data. Therefore, an efficient compression vital practical use In this paper, we propose displacement intra prediction scheme maximum two hypotheses plenoptic focused cameras. The proposed further implemented into High Efficiency Video...
Plenoptic images are one type of light field contents produced by using a combination conventional camera and an additional optical component in the form microlens arrays, which positioned front image sensor surface. This setup can capture sub-sampling with high spatial fidelity over small range, more coarsely sampled angle range. The earliest applications that leverage on plenoptic content is refocusing, non-linear distribution out-of-focus areas, SNR vs. resolution trade-offs, 3D-image...
Convolutional neural networks (CNN) represent a key application in data centers, which calls for accelerators that are: 1) efficient CNN computations; 2) having high throughput to be cost-efficient; and, 3) with adequate programming flexibility algorithm upgrades. Lacking of the availability such chip market, we designed our own. Matrix multiplication (MM) and convolution (CONV) are top-2 deep learning (DL) operations requiring intensive computation. Most existing accelerators, like GPUs...
Nocturnal water consumption (NWC), known as including stem refilling (SR) and nocturnal transpiration (NT), has been documented in many plant species, but we do not yet have a clear understanding of species differences the biotic abiotic regulation this phenomenon, especially for subtropical tropical plantations. In study, examine magnitude, seasonality, NWC, SR, NT three widely planted Eucalyptus spp., Hevea brasiliensis, Castanopsis hystrix, through measured sap meteorological variables....
The β-carbolines, mainly including harman and norharman, are a group of naturally occurring, plant-derived alkaloids, also considered as nonpolar heterocyclic aromatic amines. Sesame seed oils contain high level β-carbolines (harman norharman). In China, sesame oil blends one the most popular types vegetable blends, which can be used cooking or frying oils. Thus, it is meaningful to investigate degradation norharman) in during heating. this work, loss norharman different have been...
This study investigates the effect of involuntary motor activity paretic-spastic muscles on classification surface electromyography (EMG) signals.Two data collection sessions were designed for 8 stroke subjects to voluntarily perform 11 functional movements using their affected forearm and hand at relatively slow fast speeds. For each subject, degree present in voluntary EMG recordings was qualitatively described from such experimental protocols. Myoelectric pattern recognition analysis...
In this paper, a simple yet effective network pruning framework is proposed to simultaneously address the problems of indicator, ratio, and efficiency constraint. This paper argues that decision should depend on convolutional weights, thus proposes novel weight-dependent gates (W-Gates) learn information from filter weights obtain binary prune or keep filters automatically. To under constraints, switchable Efficiency Module constructed predict hardware latency FLOPs candidate pruned...
Rice noodles, a traditional Chinese food, are made using natural fermentation. The fermentation process plays an important role for the texture and flavour, which due to metabolism of microorganisms involved, especially lactic acid bacteria (LAB). However, rice noodles is non-controlled spontaneous process; quality products not stable. In order improve stabilize production, LAB functions on during need be elucidated. this study, community composition dynamics associated with fermented were...
Human brains are known to be capable of speeding up visual recognition repeatedly presented objects through faster memory encoding and accessing procedures on activated neurons. For the first time, we borrow distill such a capability into semantic design, namely SMTM, improve on-device CNN inference. SMTM employs hierarchical architecture leverage long-tail distribution interest, further incorporates several novel techniques put it effects: (1) encodes high-dimensional feature maps...
This paper studies one typical mobile edge computing (MEC) system, where a single user has some intensively calculating tasks to be computed by M nodes (ENs) with much more powerful capability. In particular, unmanned aerial vehicle (UAV) can act as the ENs due its flexibility and high mobility in deployment. For this we propose several EN selection criteria improve system whole performance of computation communication. Specifically, criterion I selects best based on maximizing received...
Recently, deep federated learning has attracted much attention from researchers in the fields of wireless communications, where relaying technique been shown as a powerful technology to assist signals and enhance transmission quality, which is very important development mobile edge computing (MEC) based Internet Things (IoT) networks. In relaying-aided MEC-IoT system, it vital importance deeply investigate system signal-to-noise ratio (SNR) at receiver side, mainly determines performance...
ABSTRACT The water-draw and utilization assessment (WDUA) policy is crucial for ensuring sustainable water development good governance in China. It consists of the WDUA construction projects (WDUA-CP) urban planning (WDUA-UP). Based on a systematic review policy's development, this study evaluated from perspectives text implementation effectiveness using text-mining statistical analysis methods. results show that generally successful has progressed across all regions Despite existing...
Background:This study aimed to investigate the impact of EIT-guided yoga breathing training on postoperative pulmonary complications (PPCs) for esophageal cancer patients. Material/Methods:Total 62 patients underwent radical resections cancer.Esophageal were randomized standard care group, or intervention group receiving an additional complete exercise under guidance EIT in AICU.Following extubation after esophagectomy, functions evaluated by with center ventilation (CoV), dependent silent...
With the deployment and commercialization of fifth-generation (5G) mobile communication network, access nodes data volume wireless network show a massive blowout growth trend. Taking beyond 5G (B5G) edge intelligent as research object, based on deep integration storage / computing communication, this paper focuses theory key technology system transmission, so to effectively support related applications B5G in future. This analyzes status storage, studies real field distributed system,...
Monitoring the spectrum utilization and capturing available bands is groundwork for Cognitive Radio (CR) technologies. Because management has been a politically as well technically sensitive issue, it usually very hard to access open data of real usage. In this sense, we believe grass-roots collections are important. For these be worthy investigations, need reduce several noises inserted during monitoring because diversity measurement environments. addition this, visualization must provided...
The following outlier detection problem is studied in a universal setting. Vector observations are collected each with M coordinates. When the i-th coordinate outlier, that assumed to be distributed according "outlier" distribution, distinct from common "typical" distribution governing all other Nothing known about and typical distributions except they have full supports. goal design detector best discern coordinate. A proposed shown universally exponentially consistent, single-letter...