本期目录

2021年 28卷 4期 刊出日期：2021-08-31



Artificial Intelligence

Liveness detection of occluded face based on dual-modality convolutional neural network

Ming Yue, Li Wenmin, Xu Siya, Gao Lifang, Zhang Hua, Shao Sujie, Yang Huifeng

中国邮电高校学报(英文), 2021, 28 (4). doi： 10.19682/j.cnki.1005-8885.2021.2001

摘要 ( 645 ) PDF (3096 KB)( 166 )

参考文献 | 相关文章 | 计量指标

Facial recognition has become the most common identity authentication technologies. However, problems such as uneven light and occluded faces have increased the hardness of liveness detection. Nevertheless, there are a few pieces of research on face liveness detection under occlusion conditions. This paper designs a face recognition technique suitable for different degrees of facial occlusion, which employs the facial datasets of near-infrared (NIR) images and visible (VIS) light images to examine the single-modality detection accuracy rate (experimental control group) and the corresponding high-dimensional features through the residual network (ResNet). Based on the idea of data fusion, we propose two feature fusion methods. The two methods extract and fuse the data of one and two convolutional layers from two single-modality detectors respectively. The fusion of high-dimensional features apply a new ResNet to get the dual-modality detection accuracy. And then, a new ResNet is applied to test the accuracy of dual-modality detection. The experimental results show that the dual-modality face liveness detection model improves face live detection accuracy and robustness compared with the single-modality. The fusion of two-layer features from the single-modality detector can also improve face detection accuracy by utilizing the above-mentioned dual-modality detector, and it doesn't increase the algorithm's complexity.


Resource allocation and hybrid prediction scheme for low-latency visual feedbacks to support tactile Internet multimodal perceptions

Kang Mancong, Li Xi, Ji Hong, Zhang Heli

中国邮电高校学报(英文), 2021, 28 (4). doi： 10.19682/j.cnki.1005-8885.2021.2002

摘要 ( 465 ) PDF (3476 KB)( 195 )

参考文献 | 相关文章 | 计量指标

Predicting user states in future and rendering visual feedbacks accordingly can effectively reduce the visual experienced delay in the tactile Internet (TI). However, most works omit the fact that different parts in an image may have distinct prediction requirements, based on which different prediction models can be used in the predicting process, and then it can further improve predicting quality especially under resources-limited environment. In this paper, a hybrid prediction scheme is proposed for the visual feedbacks in a typical TI scenario with mixed visuo-haptic interactions, in which haptic traffic needs sufficient wireless resources to meet its stringent communication requirement, leaving less radio resources for the visual feedback. First, the minimum required number of radio resources for haptic traffic is derived based on the haptic communication requirements, and wireless resources are allocated to the haptic and visual traffics afterwards. Then, a grouping strategy is designed based on the deep neural network (DNN) to allocate different parts from an image feedback into two groups to use different prediction models, which jointly considers the prediction deviation thresholds, latency and reliability requirements, and the bit sizes of different image parts. Simulations show that, the hybrid prediction scheme can further reduce the visual experienced delay under haptic traffic requirements compared with existing strategies.


QoE-based video segments caching strategy in urban public transportation system

Wang Hang, Li Xi, Ji Hong, Zhang Heli

中国邮电高校学报(英文), 2021, 28 (4). doi： 10.19682/j.cnki.1005-8885.2021.2003

摘要 ( 570 ) PDF (1902 KB)( 237 )

参考文献 | 相关文章 | 计量指标

With the rapid development of vehicle-based applications, entertainment videos have gained popularity for passengers on public vehicles. Therefore, how to provide high quality video service for passengers in typical public transportation scenarios is an essential problem. This paper proposes a quality of experience (QoE)-based video segments caching (QoE-VSC) strategy to guarantee the smooth watching experience of passengers. Consequently, this paper considers a jointly caching scenario where the bus provides the beginning segments of a video, and the road side unit (RSU) offers the remaining for passengers. To evaluate the effectiveness, QoE hit ratio is defined to represent the probability that the bus and RSUs jointly provide passengers with desirable video segments successfully. Furthermore, since passenger volume change will lead to different video preferences, a deep reinforcement learning (DRL) network is trained to generate the segment replacing policy on the video segments cached by the bus server. And the training target of DRL is to maximize the QoE hit ratio, thus enabling more passengers to get the required video. The simulation results prove that the proposed method has a better performance than baseline methods in terms of QoE hit ratio and cache costs.


Structural regularized twin support vector machine based on within-class scatter and between-class scatter

Wu Qing, Fu Yanlin, Fan Jiulun, Ma Tianlu

中国邮电高校学报(英文), 2021, 28 (4). doi： 10.19682/j.cnki.1005-8885.2021.2004

摘要 ( 399 ) PDF (5460 KB)( 182 )

参考文献 | 相关文章 | 计量指标

Robust minimum class variance twin support vector machine (RMCV-TWSVM) presented previously gets better classification performance than the classical TWSVM. The RMCV-TWSVM introduces the class variance matrix of positive and negative samples into the construction of two hyperplanes. However, it does not consider the total structure information of all the samples, which can substantially reduce its classification accuracy. In this paper, a new algorithm named structural regularized TWSVM based on within-class scatter and between-class scatter (WSBS-STWSVM) is put forward. The WSBS-STWSVM can make full use of the total within-class distribution information and between-class structure information of all the samples. The experimental results illustrate high classification accuracy and strong generalization ability of the proposed algorithm.


Signal processing

ESO-KELM-based minor sensor fault identification

Zhao Kai, Song Jia, Wang Xinlong

中国邮电高校学报(英文), 2021, 28 (4). doi： 10.19682/j.cnki.1005-8885.2021.2005

摘要 ( 380 ) PDF (5082 KB)( 221 )

参考文献 | 相关文章 | 计量指标

Aiming at the sensor faults of near-space hypersonic vehicles (NSHV), a fault identification method based on the extended state observer and kernel extreme learning machine (ESO-KELM) is proposed in this paper. The method is generated by a combination of the model-based method and the data-driven method. As the source of the fault diagnosis, the residual signals represent the difference between the ESO output and the result measured by the sensor in particular. The energy of the residual signals is distributed in both low frequency bands and high frequency bands. However, the energy of the sensor concentrates on the low-frequency bands. Combined with more different features detected by KELM, the proposed method devotes to improving the accuracy. Meanwhile, it is competent to calculate the magnitude of minor faults based on time-frequency analysis. Finally, the simulation is performed on the longitudinal channel of the Winged-Cone model published by the national aeronautics and space administration (NASA). Results show the validity and the accuracy in calculating the magnitude of the minor faults.


Tracking structure suitable for L6 signal

Wang Yiting, Ba Xiaohui, Liu Xueyong, Yang Ying, Li Jian, Chen Jie

中国邮电高校学报(英文), 2021, 28 (4). doi： 10.19682/j.cnki.1005-8885.2021.2006

摘要 ( 424 ) PDF (2373 KB)( 154 )

参考文献 | 相关文章 | 计量指标

A tracking structure suitable for L6 signal of quasi-zenith satellite system (QZSS) was proposed in order to track the L6 signal without other frequency assistance. Moreover, the tracking structure does not change the receiver's hardware structure. The main difference between the proposed and the traditional tracking structure lies in the generation of local codes of E, P and L branches. The method of local code generation is designed in a two-stage manner. The first stage is the generation of the P branch local code with fast fourier transform (FFT). In the second stage, the local codes of the E and L branches are obtained with the code-chip interval. The tracking structure can track and decode L6 signal separately, and track code shift keying (CSK) modulated signal as well. The structure was verified using both simulation data, generated in different conditions, and actual data obtained from QZSS satellites respectively. The results show that the improved tracking loop is able to track L6 signal without other frequency assitance. Furthermore, the biterror ratio (BER) of L6 tracking algorithm is lower than that of L1C/A assist L6 algorithm, when the Doppler remains a constant and reed solomn (RS) encode are applied. To be more specific, with the proposed structure the BER decreased by 11.40%, 17.07%, 15.00%, 11.15%, 5.19% when carrier to noise ratio (CNR) is 36 -40 dB·Hz.


Others

Cross-project software defect prediction based on multi-source data sets

Huang Junfu, Wang Yawen, Gong Yunzhan, Jin Dahai

中国邮电高校学报(英文), 2021, 28 (4). doi： 10.19682/j.cnki.1005-8885.2021.2007

摘要 ( 391 ) PDF (1116 KB)( 131 )

参考文献 | 相关文章 | 计量指标

Cross-project defect prediction (CPDP) uses one or more source projects to build a defect prediction model and applies the model to the target project. There is usually a big difference between the data distribution of the source project and the target project, which makes it difficult to construct an effective defect prediction model. In order to alleviate the problem of negative migration between the source project and the target project in CPDP, this paper proposes an integrated transfer adaptive boosting (TrAdaBoost) algorithm based on multi-source data sets (MSITrA). The algorithm uses an existing two-stage data filtering algorithm to obtain source project data related to the target project from multiple source items, and then uses the integrated TrAdaBoost algorithm proposed in the paper to build a CPDP model. The experimental results of Promise's 15 public data sets show that: 1) The cross-project software defect prediction model proposed in this paper has better performance in all tested CPDP methods; 2) In the within-project software defect prediction (WPDP) experiment, the proposed CPDP method has achieved the better experimental results than the tested WPDP method.


Research on error compensation of robot based on multi-hole measurement technology and particle swarm optimization

Wang Xianlun, Sun Yuxuan, Wang Dong, Hu Xiaowei

中国邮电高校学报(英文), 2021, 28 (4). doi： 10.19682/j.cnki.1005-8885.2021.2008

摘要 ( 342 ) PDF (2257 KB)( 118 )

参考文献 | 相关文章 | 计量指标

Robot error compensation is a technique for enhancing the positioning accuracy of the system. This paper presented an error measuring technique for serial robots based on the multi-hole measuring method, combined with the intelligent particle swarm optimisation (PSO) to obtain the optimal solution of the robot's error compensation values, thereby improving the positioning accuracy of the robot. In the experiment, the robot error was measured using self-made multi-hole measuring plates and probes, and the experimental data were combined with PSO for the error comprehensive analysis. The results showed that on this type of serial robot, the multi-hole measuring method and PSO algorithm had obvious error compensation effects, which effectively improved the positioning accuracy of the robot, with the error reduced by 35% after compensation.