WLWeiguang Li

Weiguang Li

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Weiguang Li has a total of 56 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Lightweight skeletal key point extraction algorithm based on TransPose network

    by Wenxian Zeng, Weiguang Li, Yuesong Li

    TransPose achieves excellent results in human pose estimation tasks, but it has disadvantages such as a lot of network parameters, high computational complexity of the model, and slow convergence speed. The TransPose model is improved to address the above problems. First, we use the Ghost module to change the ordinary convolutional structure in the convolutional backbone network HRNet to decrease the quantity of model parameters and calculation complexity; second, we use the multi-head mechanism of attention to increase the encoder learnable information and improve the model accuracy; third, we introduce the normalization structure of the front layer to accelerate the model convergence. It is proved by experimental results that the amount of computation of the improved TransPose model on COCO dataset is reduced from 21.8G to 15G, the number of parameters is reduced from 17.5M to 4.85M, and the model converges faster while maintaining a high accuracy rate.

  • Open Access

    Controlling the physical field using the shape function technique

    by ThanhTrung Trang, ThanhLong Pham, Yueming Hu, Weiguang Li, Shoujin Lin

    A field is described as a region under the influence of some physical force, such as electricity, magnetism, or heat. It is a continuous distribution in the space of continuous quantities. The characteristics of the field are that the values vary continuously between neighboring points. However, because of the continuous nature of the field, it is possible to approximate a physical field of interpolation operations to reduce the cost of sampling and simplify the calculation. This article introduces the modeling of the parametric intensity of physical fields in a general form based on the interpolation shape function technique. Besides the node points with sample data, there are interpolation points, whose accuracy depends significantly on the type of interpolation function and the number of node points sampled. Therefore, a comparative analysis of theoretical shape functions (TSFs) and experimental shape functions (ESFs) is carried out to choose a more suitable type of shape function when interpolating. Specifically, the temperature field is the quantity selected to apply, analyze, and conduct experiments. Theoretical computations, experiments, and comparisons of results have been obtained for each type of shape function in the same physical model under the same experimental conditions. The results show that ESF has an accuracy (error of 0.66%) much better than TSF (error of 10.34%). Moreover, the field model surveyed by a generalized reduced gradient algorithm allows for identifying points with the required parameter values presented in detail. The illustrated calculations on temperature field control in the article show that the solution for both forward and reverse problems can be determined very quickly with high accuracy and stability. Therefore, this technique is expected to be entirely feasible when applied to thermal control processes such as drying in paint technology, kilns, and heat dissipation in practice.

  • Open Access

    Modelling and graphing the Wi-Fi wave field using the shape function

    by ThanhTrung Trang, ThanhLong Pham, Yueming Hu, Weiguang Li, Shoujin Lin

    This article deals with modelling the Wi-Fi wave field, which is not visible to the naked eye, into a mathematical structure for assigning colours to observe the wave strength change compared to the source layout’s location correlation of obstacles in the survey environment. The shape function is used as a tool in modelling combined with finite sample measurement data to determine the basic parameters of the shape function. These will allow the wave field to be modelled using two techniques: the theoretical shape function and the experimental shape function. The measuring tool used in the article is the Wi-Fi Analyzer App that runs on a smartphone. The colour-assigned representation of an invisible field such as a wave will help select the signal maximal or minimization positions most beneficial to the user. The proposal is easily applied in the wave field and other quantities such as temperature, light, and concentration with similar techniques.


  • Nonclassical Ion Channels in Learning and Memory

    by Ze-Jie Lin, Xue Gu, Tian-Le Xu, Wei-Guang Li

    The well-documented involvement of voltage- and ligand-gated ion channels in the control of intrinsic excitability and synaptic plasticity and thereby in learning and memory has greatly deepened the understanding of the molecular basis of brain function, and some of these ion channel knowledges have been successfully applied in clinical therapeutics. However, less is known about the recently discovered ion channels, including acid-sensing ion channels (ASICs), transient receptor potential (TRP) channels, and mechanically activated channels (MACs), which are collectively referred to as nonclassical ion channels, especially their roles in learning and memory. This chapter thus discusses the emerging evidence for the engagement of ASICs as paradigmatic examples for the nonclassical ion channels in learning and memory, and reviews the biologically salient features of nonclassical ion channels. We highlight the more specialized activities necessary for channel function in particular cellular contexts under the behavioral tasks of learning and memory, in hopes of advancing the mechanistic understanding of cellular and molecular substrates in the memory field.

  • Open Access

    Feature frequency extraction algorithm based on matching pursuit and its application in axis orbit

    by Weiguang Li, Xinxin Zhang, Zhen Li, Jian Huang, Wenwei Cai

    Rotor displacement signal is an important state characteristic parameter of rotating machinery, which contains abundant operation information of rotating machinery. Aiming at the problem of purifying axis orbit in large -scale rotor test-bed, a new method of extracting the characteristic frequency and synthesizing axis orbit is proposed based on the matching pursuit algorithm (MP). In this paper, the sparse decomposition of signals in the algorithm is implemented by MP algorithm. The algorithm can accurately extract single or multiple feature frequencies. The purified curve is clearer and more efficient. Fault types can be found by extracting different characteristic frequency synthetic axis trajectories, so as to understand the operation status of the large-scale rotor test-bed.

  • Feature frequency extraction based on singular value decomposition and its application on rotor faults diagnosis

    by Zhen Li, Weiguang Li, Xuezhi Zhao

    The selection of effective singular values using the singular value decomposition (SVD) method has always been a hot topic. In this paper, we found that there was a special relationship between effective singular values and feature frequency components. Theoretical derivations illustrated that each frequency component produced two adjacent nonzero singular values with one ranking another closely. Size of singular values was directly proportional to amplitude of feature frequency. The number of singular values was only related to the number of feature frequency components. For these discoveries, a novel feature frequency separation method based on SVD was proposed, through which axis orbits of large rotating machines were readily purified. The results show that the algorithm was very accurate in feature frequency extraction.


  • Open Access

    Development of a New Robotic Ankle Rehabilitation Platform for Hemiplegic Patients after Stroke

    by Quanquan Liu, Chunbao Wang, Jian Jun Long, Tongyang Sun, Lihong Duan, Xin Zhang, Bo Zhang, Yajing Shen, Wanfeng Shang, Zhuohua Lin, et al.

    A large amount of hemiplegic survivors are suffering from motor impairment. Ankle rehabilitation exercises act an important role in recovering patients’ walking ability after stroke. Currently, patients mainly perform ankle exercise to reobtain range of motion (ROM) and strength of the ankle joint under a therapist’s assistance by manual operation. However, therapists suffer from high work intensity, and most of the existed rehabilitation devices focus on ankle functional training and ignore the importance of neurological rehabilitation in the early hemiplegic stage. In this paper, a new robotic ankle rehabilitation platform (RARP) is proposed to assist patients in executing ankle exercise. The robotic platform consists of two three-DOF symmetric layer-stacking mechanisms, which can execute ankle internal/external rotation, dorsiflexion/plantarflexion, and inversion/eversion exercise while the rotation center of the distal zone of the robotic platform always coincides with patients’ ankle pivot center. Three exercise modes including constant-speed exercise, constant torque-impedance exercise, and awareness exercise are developed to execute ankle training corresponding to different rehabilitation stages. Experiments corresponding to these three ankle exercise modes are performed, the result demonstrated that the RARP is capable of executing ankle rehabilitation, and the novel awareness exercise mode motivates patients to proactively participate in ankle training.


  • Open Access

    Feature Frequency Extraction Based on Principal Component Analysis and Its Application in Axis Orbit

    by Zhen Li, Weiguang Li, Xuezhi Zhao

    Vibration‐based diagnosis has been employed as a powerful tool in maintaining the operating efficiency and safety for large rotating machinery. However, the extraction of malfunction features is not accurate enough by using traditional vibration signal processing techniques, owing to their intrinsic shortcomings. In this paper, the relationship between effective eigenvalues and frequency components was investigated, and a new characteristic frequency separation method based on PCA (CFSM‐PCA) was proposed. Certain feature frequency could be purified by reconstructing the specified eigenvalues. Furthermore, three significant perspectives were studied via the distribution of effective eigenvalues, and theoretical derivations were subsequently illustrated. More importantly, this proposed scheme could also be used to synthesize axis orbits of larger machines. Purified curves were so explicit and the CFSM‐PCA exhibited higher efficiency than harmonic wavelet and wavelet packet.