Current Issue

• • 2023, Vol. 46 No. 6 Published date:28 December 2023 Last issue Next issue

  • Methods and Research Progress of AI-based White Matter Tract Segmentation
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 1-0.
  • Abstract ( 521 )     　　
  • Diffusion magnetic resonance imaging allows mapping white matter fiber tracts via a process called tractography. Segmentation of white matter tracts according to fiber characteristics can assist statistical analysis and precision medicine. In this paper, we first introduce the principles of white matter tractography and segmentation. Then, we categorize state-of-the-art segmentation methods into voxel-based and fiber-based categories. Moreover, various artificial intelligence algorithms on segmentation are summarized and concluded, and an experiment is conducted to show the segmentation results. Finally, we discuss the challenges and research trends, and forecast the progress prospect of artificial intelligence in white matter tract segmentation. In summary, the review  provides a comprehensive methodological summary and diagnostic support for downstream neuroscience research in sub-healthy individuals and patients.
  • Supplementary Material | Related Articles

  • Sleep staging algorithm based on class rebalancing unsupervised domain adaptation
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 8-0.
  • Abstract ( 376 )     　　
  • Sleep disorders seriously affect human health. Deep learning automatic sleep staging algorithms based on electroencephalograms (EEG) can assist experts in accurately diagnosing patients' sleep disorders. However, the imbalance in training data hinders the learning of minority class features, and due to the differences in data distribution, the accuracy of automatic sleep staging models trained on training data often decreases when applied to real-world data. To address this issue, we propose an unsupervised domain adaptation algorithm that combines class re-balancing strategies and semi-supervised learning. In particular, a balanced loss function is introduced to mitigate data imbalance issue in sleep staging datasets. Additionally, an average teacher method is designed, and random output interpolation and related confidence thresholding are introduced to improve the accuracy of pseudo-labels. The feature distribution of target domain data is optimized through a discriminator network, thereby improving the classification accuracy on the target domain. Experiments conducted on the SHHS, Sleep-EDF, and ISRUC-Sleep datasets demonstrate the effectiveness of the proposed algorithm. Compared to direct transfer, the accuracy improves by 3.28% to 13.27%. Compared to domain statistical alignment, the accuracy is increased by 6.73% to 14.52%. Compared to adaptive domain statistical alignment methods, the accuracy is increased by 0.78% to 5.82%.
  • Supplementary Material | Related Articles

  • A Asynchronous Detection Algorithm of SSVEP-BCI Based on Maximum Posterior Criterion

  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 15-0.
  • Abstract ( 270 )     　　

  • Considering the prior probability of EEG signals is non-equal in some application scenario of brain-computer interface, we propose a asynchronous detection algorithm of steady-state visual evoked potential brain-computer interface (SSVEP-BCI) based on maximum posterior criterion. Using spatio-temporal equalization multi-window technology, a dynamic window algorithm based on sequential detection is designed by introducing maximum posterior criterion, which makes full use of prior information to improve the performance of brain-computer interface systems. Experiments show that compared with the traditional spatio-temporal equalization multi-window algorithm, the target recognition accuracy of the proposed algorithm is increased, and the average practical bit rate has a substantial improvement. Besides, the instruction time and the average false alarm rate are significantly reduced. Thus the effectiveness and practicability of the proposed algorithm are verified.

  • Supplementary Material | Related Articles

  • Incompressible Number Density Based SPH Model for Simulation of Silicone Oil Tamponade and Emulsification
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 20-0.
  • Abstract ( 209 )     　　
  • Due to the lack of quantitative research on the appropriate amount of silicone oil and the removal time, pars plana vitrectomy combined with silicone oil tamponade is prone to various ocular complications that affect postoperative visual recovery. To address this practical issue, a smooth particle hydrodynamics (SPH) method based on incompressible number density is proposed to simulate the coupled interaction of silicone oil and water inside the eyeball. Specifically, the surface tension model and local equilibrium model are combined to simulate the strong surface tension and emulsification diffusion phenomenon of silicone oil, and the elastic model is introduced to achieve stable coupling between the silicone oil and intraocular fluid with the ocular tissue. The experimental results show that the visualization simulation of silicone oil tamponade and emulsification process can provide an effective computer-aided means for doctors’ clinical diagnosis, surgical planning, and prognosis monitoring, so as to improve the success rate of surgery and postoperative outcomes.
  • Supplementary Material | Related Articles

  • Image Segmentation Based on Federal Style Transfer
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 27-0.
  • Abstract ( 414 )     　　
  • In this work, we propose an image segmentation method based on federated style transfer to solve the non-independent and identically distributed (non-IID) problem in federated learning. By sharing style information that is not sensitive to user privacy, this method generates synthetic data for data expansion and reduces data differences between different users while ensuring that important structural information of data is not disclosed. Experiment results show that this method effectively alleviates the influence of non-IID problem among nodes on the performance of the federated model in the liver image segmentation task. Therefore, the proposed method can further improve the performance of federal model, which provides the possibility to break the data island and establish a general model in medical field.
  • Supplementary Material | Related Articles

  • Object Localization Method in Microscopic Image Based on Shape Perception
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 33-0.
  • Abstract ( 206 )     　　
  • To deal with the inaccurate positioning issue of the needle tip in Intracytoplasmic sperm injection microscopic images, we propose a shape-sensing-based microscopic image needle tip positioning method. In particular, the target shape and boundary information are proposed as constraints on the loss function of the image segmentation network, driving the convolutional neural network to pay more attention to the shape and boundary features of the target during the training process. The experimental results show that the performance of the proposed loss function is increased by more than 7% compared to the three commonly used baseline segmentation models. Thus, the method can be used as a plug-and-play module to improve the generalization ability of the algorithm. In addition, its localization accuracy exceeds the state-of-the-art methods in the field.
  • Supplementary Material | Related Articles

  • An Interpretable Prediction Model for Heart Disease Risk Based on Improved Whale Optimized LightGBM
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 39-0.
  • Abstract ( 282 )     　　
  • Aiming at the problems of low accuracy and poor interpretability of existing heart disease risk prediction models, an interpretable heart disease risk prediction model based on improved whale optimized light gradient boosting machine (LightGBM) is proposed. First, the deep auto-encoder is used to effectively reduce the data dimensionality. Then, the whale optimization algorithm is improved by various strategies to obtain the global optimal solution of LightGBM hyper-parameter, including halton sequence initialization population, nonlinear convergence factor and dynamic spiral update. Finally, the important features of the proposed model are explained and analyzed by the shapley additive explanations method. Compared with other mainstream dimensionality reduction methods and classification models, experimental results show that the proposed model can obtain higher prediction accuracy and can efficiently extract the potential characteristics of heart disease risk factors.
  • Supplementary Material | Related Articles

  • Automotive In-Vehicle Time-Sensitive Networking: the State of the Art and Prospect
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 46-0.
  • Abstract ( 741 )     　　
  • In order to facilitate researchers in both network and automotive fields to have a comprehensive understanding of the automotive in-vehicle time-sensitive networking, this paper summarizes the current research status of the automotive in-vehicle time-sensitive networking, including key technologies such as time synchronization, traffic scheduling, reliability and security, and resource management, as well as application scenarios of electrical and electronic architecture, integrated design, and external vehicle communications.Challenges faced by in-vehicle time-sensitive networking are analyzed, which are mainly reflected in the strict communication performance requirements of electronic and electrical architecture, complex multi-stream hybrid scheduling, difficulties in integrating time-sensitive networking and the fifth generation of mobile communications system-Internet of Vehicles, and imperfect tool chain. Based on these challenges, the development trend of in-vehicle time-sensitive networks is prospected.
  • Supplementary Material | Related Articles

  • Communication-Sensing Integrated Resource Allocation Algorithm in Vehicular Networks
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 55-0.
  • Abstract ( 457 )     　　
  • With the help of millimeter wave communication technology, the efficient transmission of sensing information between different vehicles can be realized, thus enhancing the sensing ability of the vehicles. In this paper, a time-division integrated sensing and communication system is adopted for extended sensor scenarios in enhanced applications of Internet of Vehicles. The matching problem between communication vehicles and the time resource allocation problem in the integrated sensing and communication system are studied with the radar mutual information rate as a performance evaluation indicator. To solve the problems, an optimization algorithm based on the Hungarian algorithm and simulated annealing is proposed. Simulation results show that the proposed algorithm can effectively extend the sensing range of vehicles and improve the overall sensing performance of vehicle networks.
  • Supplementary Material | Related Articles

  • Fuzzy logic Decision Multiple Access and Power Allocation Scheme for Indoor Visible Light Communication
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 61-0.
  • Abstract ( 221 )     　　
  • In order to meet the needs of large-scale user access and differentiated user request rates, non-orthogonal multiple access (NOMA) and orthogonal multiple access schemes for fuzzy logic decision-making in indoor visible light communications (VLC) are studied. The proposed scheme considers the number of users, user request rate and user channel gain, and designs a mode selection algorithm based on fuzzy logic to decide the user access NOMA or orthogonal multiple access mode. In power allocation phase, an unequal subcarrier power allocation between NOMA group and power allocation between users in a NOMA group according to user’s differentiated request rates are designed to improve the throughput of the system. The simulation results show that proposed scheme can improve the VLC system throughput and average user satisfaction.
  • Supplementary Material | Related Articles

  • A Self-distillation Lightweight Image Classification Network Scheme
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 66-0.
  • Abstract ( 285 )     　　
  • Image classification tasks often compress neural network models to reduce the number of parameters, which will lead to a decrease in classification accuracy. In order to solve this problem, a self-distillation lightweight image classification network scheme is proposed. First, a lightweight attention module with negligible calculation and parameter amounts is introduced into the self-distillation framework to reduce the parameter amount and calculation amount of the self-distillation framework, thereby achieving a lightweight self-distillation framework. Then, group convolution and depthwise separable convolution are used to compress the residual network and VGG11 network. Next, two compressed neural networks are adopted as teacher model. According to depth of the teacher model, multiple shallow classifiers as student models are constructed. Finally, a lightweight self-distillation framework is built. Experimental results show that the proposed scheme not only ensures original self-distillation effect, but also greatly reduces the number of parameters of the compressed image classification network without being lower than the accuracy of the original classification, and reduces the difficulty of model deployment.
  • Supplementary Material | Related Articles

  • Chaotic Spread Spectrum Communication with Coupled Map Lattice based on Cellular Automata
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 72-0.
  • Abstract ( 155 )     　　
  • In order to solve the problem of uneven output distribution of chaotic systems and the degradation of chaotic system dynamics under the limited precision, a pseudo-random coupled map lattice chaotic spreading code generation algorithm is proposed based on the iterative rules of elementary cellular automata and coupled map lattices. First, elementary cellular automata is used to control the coupling objects of each lattice in coupled map lattices. The initial iteration value is generated by the XOR operation of different elementary cellular automata iteration rules. The index is carried out according to the elementary cellular automata iteration results, and the system index rules are designed. Then, the coupling coefficient generation algorithm is designed, and the coupling coefficient value is obtained according to the elementary cellular automata initial value and two secret integers. The lattice index in the coupled map lattices is mapped to the cell index in the elementary cellular automata, and then the corresponding disturbance values and disturbance symbols are returned according to the algorithm rules, which can effectively weaken the degradation of the dynamic characteristics of the chaotic system and enhance the chaotic characteristics of the spatiotemporal chaotic system. Finally, the scheme of chaotic spread spectrum communication system based on pseudo-random coupled image lattice is designed. The pseudo-random coupled map lattice chaotic spread spectrum code is directly spread and modulated by BPSK. The simulation results show that the chaotic spread spectrum code has passed the NIST test and has excellent pseudo-randomness. The proposed scheme has good bit error rate performance and has extensive application prospects in the field of secure communication.
  • Supplementary Material | Related Articles

  • Research on Service Resource Allocation Algorithm for 5G-SBA
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 77-0.
  • Abstract ( 214 )     　　
  • Based on the network function virtualization and service-based architecture (SBA) in the core network of fifth generation of mobile communications system (5G), a multi-policy service resource scheduling algorithm based on Viterbi is proposed. First, by modeling the underlying network and server nodes, the optimization goal is transformed into the problem of minimizing the scheduling cost of virtualized network function. Then, the algorithm evaluates factors such as slice type, network function type, node load rate, and proposes three sub-strategies in different scenarios to screen candidate nodes. Finally, combined with the Vertibi algorithm, the hidden Markov chain model is constructed, and the target nodes and paths scheduled by the service function chain are output, and the underlying resources held by the virtualized network function are occupied based on this.  Simulation results show that, compared with the conventional resource load optimization algorithm and random scheduling algorithm, the proposed algorithm has better performance in terms of service scheduling cost, network resource utilization, average overhead and average delay.
  • Supplementary Material | Related Articles

  • A Clustering Routing Algorithm Based on Improved Genetic Algorithm
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 83-0.
  • Abstract ( 259 )     　　
  • To overcome the deficiency of existing clustering algorithms of wireless sensor networks, a clustering routing algorithm based on improved genetic algorithm is proposed. First, to accelerate convergence and avoid local convergence of traditional genetic algorithm, an improved genetic algorithm based on harmony algorithm and adaptive optimization method is proposed. Then, the optimal number of cluster heads is derived using the network energy consumption model and node distribution model. Finally, the optimal cluster heads are selected using improved genetic algorithm. The energy of the node, the distance from the sink, the density of neighbor nodes and other factors are considered during the design of the fitness function. In order to balance and reduce energy consumption, the energy and location factors are considered when defining the clustering selection function and relay cost function. Simulation results show that the proposed algorithm can realize load balancing and  reduce network energy consumption effectively.
  • Supplementary Material | Related Articles

  • A High Accuracy Time Synchronization Method for 5G-TSN System Based on Resident Time Compensation
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 89-0.
  • Abstract ( 456 )     　　
  • In order to study the time synchronization of time sensitive networking (TSN) in 5th generation mobile communication (5G) and TSN converged network deeply, TSN translating network elements are designed and developed based on the OMNeT++ platform, and simulation experiments are carried out. A delay measurement method based on compensation of 5G system resident time is proposed to solve the problem that the measured value of the transmission delay of 5G system has a large deviation from the actual value. The simulation results show that if the 5G system and the TSN system have the same and high time synchronization accuracy, the improved delay measurement method enables microsecond level time synchronization of TSN in the converged network.
  • Supplementary Material | Related Articles

  • Miao Costume Image Segmentation Algorithm Generalized Enhanced Interval Type-2 FCM
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 95-0.
  • Abstract ( 192 )     　　
  • For the problem of increased difficulty in segmentation due to the colorful, rich shape, complex texture and other characteristics of Miao costume patterns,  a novel method of interval type-2 fuzzy set is proposed, Miao costume image segmentation algorithm generalized enhanced interval type-2 fuzzy c-means (FCM). Interval type-2 fuzzy sets can effectively deal with the uncertainty and fuzziness in the process of image segmentation, and can grasp more image details. The proposed algorithm that adds a competitive penalty term to the interval type-2 FCM, can obtain high robustness and convergence. Enhanced KM algorithm is used to type-reduction and defuzzification of optimizing the centroid so as to speed up the model. The experimental results show that the proposed algorithm can achieve more accurate segmentation  than that of other algorithms in different datasets.
  • Supplementary Material | Related Articles

  • Environmental Sound Classification Based on Compact Bilinear Attention Network
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 102-0.
  • Abstract ( 207 )     　　
  • Local regional differences can make it difficult to classify environmental sounds accurately. Therefore, an environmental sound classification based on compact bilinear attention network is proposed. First, multi-dimensional time-frequency features are introduced to fully characterize the characteristics of environmental sound. Second, online random erasing data augmentation is introduced to avoid overfitting of the trained model due to lack of dataset and improve sample diversity. Finally, with the unchanged compact bilinear network framework, DensNet-169 is adopted as the feature extraction module, and the channel spatial location attention module is introduced to pay attention to the differences of local regions of environmental sound features. The experimental results show that the accuracy of the proposed method on ESC-10 and ESC-50 datasets can reach 96.0% and 87.9%, respectively, both of which are better than human ear recognition accuracy.
  • Supplementary Material | Related Articles

  • Anonymous Zone Construction Scheme Combining Smart Contract and MADM Credit Evaluation
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 108-0.
  • Abstract ( 195 )     　　
  • Aiming at the trust problem among users in the construction process of distributed k-anonymity technology, an anonymous zone construction scheme combining credit evaluation and smart contracts is proposed. First, multiple attributes involved in the anonymous zone construction process are introduced to quantify the user's credit indicators in the anonymous zone construction and convert them into credit values through a multi-attribute decision making (MADM). Second, combined with blockchain technology, a smart contract function based on credit value evaluation is designed. Finally, the smart contract function verifies the user's identity information and automatically calculates the user's reputation value, and requests the user and collaborative users to only interact through the smart contract to complete the construction of the anonymous area. The experimental results show that the proposed scheme has lower communication overhead and calculation delay, which ensures the generation efficiency of anonymous regions and provides better protection for users' location privacy.
  • Supplementary Material | Related Articles

  • Temperature Compensation of Ice Measurement System Based on Improved Particle Swarm Optimization Algorithm
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 115-0.
  • Abstract ( 170 )     　　
  • The measurement value of the icing sensor designed with three-terminal piezoelectric ceramics is greatly affected by temperature. In order to improve the measurement accuracy of the sensor, an adaptive simulated annealing particle swarm optimization algorithm is proposed to compensate for temperature. This algorithm integrates the simulated annealing algorithm and the particle swarm algorithm, uses the nonlinear hyperbolic tangent function to control the change of the inertia weight coefficient, and uses the linear change strategy to control the values of the social learning factor and the self-learning factor, so that the optimization focus varies at different stages. The inertia weight coefficient , the social learning factor and the self-learning factor are adaptively changed, which solves the problem that the algorithm is easy to fall into the local optimal solution. The experimental results show that after the temperature compensation of this algorithm, the measurement error of the freezing state is reduced to about 3% ,  and the measurement accuracy is improved to 0.1mm, which indicates that the compensation algorithm can effectively reduce the influence of -5 ℃~-55 ℃ on the measurement results of the sensor and improve the measurement accuracy.
  • Supplementary Material | Related Articles

  • An Improved Variable Step Size Least Mean Square Adaptive Filtering Algorithm Based on RBFNN
  • Journal of Beijing University of Posts and Telecommunications. 2023, 46(6): 121-0.
  • Abstract ( 270 )     　　
  • In order to further improve the convergence and stability of Radial basis function neural network nonlinear adaptive filtering, an improved variable step size least mean square adaptive filtering algorithm based on Radial basis function neural network (RBFNN) was proposed. On the basis of the variable step size least mean square algorithm and the inverse hyperbolic tangent function, a variable scale function is used to replace the fixed parameters in the variable step size model in order to solve the drawbacks of fixed step size in the algorithm and the problem of selecting fixed parameters in the variable step size model. Then, the improved algorithm is applied to RBFNN to update and train the center, width, and output weight parameters of the network, in order to improve the performance of RBFNN filtering under gradient algorithm. Finally, the simulation comparison experiments are carried out in nonlinear system identification and chaotic time series prediction. The results show that the proposed algorithm has obvious advantages in terms of convergence rate and steady-state error performance.
  • Supplementary Material | Related Articles