GCGuanghua Chen

Guanghua Chen

*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


Guanghua Chen has a total of 53 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Finger Vein Recognition Based on ResNet With Self-Attention

    by Zhibo Zhang, Guanghua Chen, Weifeng Zhang, Huiyang Wang

    To solve the problem of low accuracy and high computational resource consumption in finger vein recognition, a finger vein recognition model based on ResNet with self-attention (FV-RSA) is proposed. This model combines global focusing ability of self-attention mechanism and local feature extraction ability of CNN, which improves recognition accuracy. To reduce the number of parameters and floating-point operations, self-attention and convolution share linear projections by pointwise convolution. Self-attention and CNN are fused in the convolution and self-attention (CASA) block connected by skip connection to avoid gradient vanishing or gradient exploding. During the training phase, we use a variable learning rate with cosine annealing to avoid falling into local optimum. Experiments show that the method works well on the public database, which can not only improve the accuracy, but also reduce the number of parameters and computational complexity.


  • Open Access

    Association of China’s two-child policy with changes in number of births and birth defects rate, 2008–2017

    by Hanyi Chen, Ting Wei, Haiyin Wang, Yi Zhou, Hua Chen, Lianghong Sun, Shaotan Xiao, Wuren Ma, Huijuan Zhao, Guanghua Chen, et al.

    Background: In October 2015, China’s one-child policy was universally replaced by a so-called two-child policy. This study investigated the association between the enactment of the new policy and changes in the number of births, and health-related birth outcomes.Methods: We used difference-in-difference model to analyse the birth record data in Pudong New Area, Shanghai.The design is descriptive before-and-after comparative study.Results: The data covered three policy periods: the one-child policy period (January 2008 to November 2014); the partial two-child policy period (December 2014 to June 2016); the universal two-child policy period (July 2016 to December 2017). There was an estimate of 7656 additional births during the 18 months of the implementation of the universal two-child policy. The trend of monthly percentage of births to mothers aged ≥35 increased by 0.24 percentage points (95% confidence interval 0.19 to 0.28,p < 0.001) during the same period. Being a baby boy, preterm birth, low birth weight, parents with lower educational attainment, and assisted delivery were associated with a higher risk of birth defects.Conclusions: The universal two-child policy was associated with an increase in the number of births and maternal age. Preterm birth, low birth weight, and assisted delivery were associated with a higher risk of birth defects, which suggested that these infants needed additional attention in the future.


  • A Driving Comfort Evaluation Method Based on Smartphone Testing Data

    by Guanghua Chen, Jinxi Zhang

    The driving vibration caused by the pavement roughness is an important factor affecting the driving comfort, and the driving vibration is speed-dependent, so there are some deficiencies in the method of reflecting the driving feeling only by the pavement roughness index. Small smart devices represented by smartphone can detect vibration and speed. In order to provide a new reference method for pavement driving comfort evaluation, Honda-CRV and smartphone were used as the experimental vehicle and data collection equipment. Driving experiments were carried out on several ordinary asphalt pavement sections with different smoothness, and the driving vibration and speed data were obtained by experiment phone. Taking the vector synthetic vibration acceleration as the vibration index, a new index for pavement driving comfort evaluation -- comfort rate (CR) was put forward from the point of view of the time of road section driving vibration process. The CRs of different experiment sections were calculated, the relationship between CR and riding quality index (RQI) was analyzed, the feasibility of using CR as the evaluation index of driving comfort was discussed. Finally, according to the CR distribution range of the experimental sections and the meaning of CR, the CR rating method for driving comfort evaluation was proposed, which provided a new way to evaluate the driving comfort with driving intelligent perception data.

  • Finger Vein Recognition Based on Multi-Receptive Field Bilinear Convolutional Neural Network

    by Kaixuan Wang, Guanghua Chen, Hongjia Chu

    In order to solve the problems of many model parameters during training, long training time and low recognition accuracy, a finger vein recognition method was proposed that uses multi-receptive field bilinear convolutional neural network (MRFBCNN) . The network can obtain the second-order features of finger veins, and better distinguish finger veins with small differences between classifications. Then the lightweight neural network is used to reduce network parameters and computational complexity, so that the algorithm can be easily implemented in the hardware with limited resources. Finally, a dimensional interactive attention mechanism (DIAM) is designed to enhance the correlation between channels and spaces, and further improve the recognition accuracy. Experimental results show that this method not only improves the accuracy of finger vein recognition, but also reduces the training time and model parameters, so it is suitable for the practical application of finger vein recognition.


  • Open Access

    Implementation of the NewHope Protocol for Post-Quantum Cryptography

    by Zhenghao Lin, Guanghua Chen

    In this paper, we propose a faster and less resource intensive implementation of the NewHope protocol to address the problems of resource-intensive random number generators and slow running NTT modules in NewHope implementations. In the random number generation module, choose the lightweight pseudo-random number generator (PRNG) Trivium instead of the SHAKE function to reduce the resource usage in the random number generation module. In the NTT transformation module, a pipeline structure is used in combination with a parallel structure of four butterfly units to shorten the cycle time of butterfly operations and achieve acceleration. The experimental results show that, compared with Kuo’s design, the overall time for key exchange in this paper is reduced by 25.3%, and the consumption of FFs and LUTs is reduced by 20.2% and 38.0%, respectively.


  • Image-set based face recognition using K-SVD dictionary learning

    by Jingjing Liu, Wanquan Liu, Shiwei Ma, Meixi Wang, Ling Li, Guanghua Chen


  • Open Access

    A Kinetic Model for the Spread of Rumor in Emergencies

    by Guanghua Chen, Huizhang Shen, Teng Ye, Guangming Chen, Naphtali Kerr

    To study rumor spreading, a novel rumor spreading model with latent, constant recruitment, and varying total population is investigated. We carry out a global qualitative analysis of the rumor spreading model with latent, constant recruitment, and varying total population and studied the existence and stability of the rumor-elimination and prevailing equilibria and the impact different management strategies have on the rumor spreading process. The Runge-Kutta method is used for numerical simulation, and the simulation results imply that increasing the efforts to crack down on rumor spreaders, the timely disclosure of transparent information, the popularization of scientific knowledge, and the improvement of the ignorants discrimination may be effective rumor control measures.


  • The Psychological Impact of Exposure to the 2008 Snowstorms on Migrant Workers in China

    by Guanghua Chen, Huizhang Shen, Guangming Chen, Naphtali Kerr, Jun Zhao

    This study aimed to estimate the prevalence of posttraumatic stress disorder (PTSD) and to identify associated risk factors among migrant workers 4 years after the 2008 snowstorm in China. A cross-sectional survey of 528 respondents was conducted to collect data in Changsha, the capital of Hunan Province. A multistage cluster sampling method was applied in selecting subjects from migrant workers’ gathering areas. Association between the independent variables and PTSD was analyzed using logistic regression analysis. The prevalence of PTSD was 17.42% among migrant workers in Changsha. Being male, being older than 35 years, being an electrician, having dependent children, suffering property damage, being without medical insurance, and having low social support were risk factors significantly related to the development of PTSD. Effective and sustainable mental health services should be directed particularly to the migrant workers, who are among the groups most vulnerable to the direct impact of snowstorms.