中国邮电高校学报(英文版), 2022, 29 (2). doi： 10. 19682/ j. cnki. 1005-8885. 2022. 0011
In order to study the role of the new technological concept of shared experiences in the digital interactive experience of cultural heritage and apply it to the digital interactive experience of cultural heritage to solve the current problems in this field, starting from the mixed reality (MR) technology that the shared experiences rely on, proper software and hardware platforms were investigated and selected, a universal shared experiences solution was designed, and an experimental project based on the proposed solution was made to verify its feasibility. In the end, a proven and workable shared experiences solution was obtained. This solution included a proposed MR spatial alignment method, and it integrated the existing MR content production process and standard network synchronization functions. Furthermore, it is concluded that the introduction and reasonable use of new technologies can help the development of the digital interactive experience of cultural heritage. The shared experiences solution for the digital interactive experience of cultural heritage balances investment issues in the exhibition, display effect, and user experience. It can speed up the promotion of cultural heritage and bring the vitality of MR technology to relevant projects.
陶艺峰 宋宇程 徐梦秋 张闯 吴铭 白苏乐
中国邮电高校学报(英文版), 2022, 29 (2). doi： 10. 19682/ j. cnki. 1005-8885. 2022. 0012
In the long history of more than 1 500 years, Dunhuang murals suffered from various deteriorations causing irreversible damage such as falling off, fading, and so on. However, the existing Dunhuang mural restoration methods are time-consuming and not feasible to facilitate cultural issemination and permanent inheritance. Inspired by cultural computing using artificial intelligence, gated-convolution-based dehaze net (GD-Net) was proposed for Dunhuang mural refurbishment and comprehensive protection. First, a neural network with gated convolution was applied to restore the falling off areas of the mural to ensure the integrity of the mural content. Second, a dehaze network was applied to enhance image quality to cope with the fading of the mural. Besides, a Dunhuang mural dataset was presented to meet the needs of deep learning approach, containing 1 180 images from the Cave 290 and Cave 112 of the Mogao Grottoes. The experimental results demonstrate the effectiveness and superiority of GD-Net.
黄佩 张萌 万柳 雷轩铮
中国邮电高校学报(英文版), 2022, 29 (2). doi： 10. 19682/ j. cnki. 1005-8885. 2022. 0013
In the context of interdisciplinary research, using computer technology to further mine keywords in cultural texts and carry out semantic analysis can deepen the understanding of texts, and provide quantitative support and evidence for humanistic studies. Based on the novel A Dream of Red Mansions, the automatic extraction and classification of those sentiment terms in it were realized, and detailed analysis of large-scale sentiment terms was carried out. Bidirectional encoder representation from transformers (BERT) pretraining and fine-tuning model was used to construct the sentiment classifier of A Dream of Red Mansions. Sentiment terms of A Dream of Red Mansions are divided into eight sentimental categories, and the relevant people in sentences are extracted according to specific rules. It also tries to visually display the sentimental interactions between Twelve Girls of Jinling and Jia Baoyu along with the development of the episode. The overall F1 score of BERT-based sentiment classifier reached 84-89%. The best single sentiment score reached 91-15%. Experimental results show that the classifier can satisfactorily classify the text of A Dream of Red Mansions, and the text classification and interactional analysis results can be mutually verified with the text interpretation of A dream of Red Mansions by literature experts.
陈佳舟 Amal Ahmed Hasan Mohammed 黄可妤 缪永伟
中国邮电高校学报(英文版), 2022, 29 (2). doi： 10. 19682/ j. cnki. 1005-8885. 2022. 0014
In the recent decade, many approaches of rough line drawing simplification were proposed, but they are not well summarized yet, especially from the perspective of Chinese cultural computing. In this paper, a comprehensive review of existing line drawing simplification methods was presented, including their algorithms, advantages/ disadvantages, inputs/ outputs, datasets and source codes, etc. For raster line drawings, related implification work was discussed according to four main categories: fitting-based methods, tracing-based methods, field-based methods, and learning-based methods. For vector line drawings, a deep investigation was introduced for two major steps of simplification: stroke grouping and stroke merging. Finally, conclusions were given, key challenges and future directions of line drawing simplification for Chinese traditional art were thoroughly discussed.
中国邮电高校学报(英文版), 2022, 29 (2). doi： 10. 19682/ j. cnki. 1005-8885. 2022. 0015
An Avatar-like robot in a virtual museum environment was designed to perform the function of telepresence and teleoperation, and make the three-dimensional (3D) effect through a binocular camera and a virtual reality (VR) head-mounted display (HMD). This robot supports users to participate in the exhibition remotely in a new and interactive way in multiple scenarios. The results show that the system has good usability and is worth further optimizing.
中国邮电高校学报(英文版), 2022, 29 (2). doi： 10. 19682/ j. cnki. 1005-8885. 2022. 0001
Aiming to solve the poor performance of low illumination enhancement algorithms on uneven illumination images, a low-light image enhancement (LIME) algorithm based on a residual network was proposed. The algorithm constructs a deep network that uses residual modules to extract image feature information and semantic modules to extract image semantic information from different levels. Moreover, a composite loss function was also designed for the process of low illumination image enhancement, which dynamically evaluated the loss of an enhanced image from three factors of color, structure, and gradient. It ensures that the model can correctly enhance the image features according to the image semantics, so that the enhancement results are more in line with the human visual experience. Experimental results show that compared with the state-of-the-art algorithms, the semantic-driven residual low-light network (SRLLN) can effectively improve the quality of low illumination images, and achieve better subjective and objective evaluation indexes on different types of images.
潘晓英 魏苗 王昊 贾丰竹
中国邮电高校学报(英文版), 2022, 29 (2). doi： 10.19682/j.cnki.1005-8885.2022.0004
The sensing light source of the line scan camera cannot be fully exposed in a low light environment due to the extremely small number of photons and high noise, which leads to a reduction in image quality. A multi-scale fusion residual encoder-decoder (FRED) was proposed to solve the problem. By directly learning the end-to-end mapping between light and dark images, FRED can enhance the image's brightness with the details and colors of the original image fully restored. A residual block (RB) was added to the network structure to increase feature diversity and speed up network training. Moreover, the addition of a dense context feature aggregation module (DCFAM) made up for the deficiency of spatial information in the deep network by aggregating the context's global multi-scale features. The experimental results show that the FRED is superior to most other algorithms in visual effect and quantitative evaluation of peak signa-to-noise ratio (PSNR) and structural similarity index measure (SSIM). For the factor that FRED can restore the brightness of images while representing the edge and color of the image effectively, a satisfactory visual quality is obtained under the enhancement of low-light.
卢莹 黄世奇 王文庆 孙柯
中国邮电高校学报(英文版), 2022, 29 (2). doi： 10. 19682/ j. cnki. 1005-8885. 2022. 0003
The physical principle of infrared imaging leads to the low contrast of the whole image, the blurring of contour and edge details, and it is also sensitive to noise. To improve the quality of infrared image and visual effect, an adaptive weighted guided filter (AWGF) for infrared image enhancement algorithm was proposed. The core idea of AWGF algorithm is to propose an adaptive strategy to update the weights of guided filter (GF) parameters, which not only improves the accuracy of regularization parameter estimation in GF theory, but also achieves the purpose of removing infrared image noise and improving its detail contrast. A large number of real infrared images were used to verify AWGF algorithm, and good experimental results were obtained. Compared with other guided filtering algorithms, the halo phenomenon at the edge of infrared images processed by the AWGF algorithm is significantly avoided, and the evaluation parameter values of information entropy (IE), average gradient (AG), and moment of inertia (MI)are relatively high. This shows that the quality of infrared image processed by the AWGF algorithm is better.
中国邮电高校学报(英文版), 2022, 29 (2). doi： 10. 19682/ j. cnki. 1005-8885. 2022. 0002
To realize the distributed storage and management of a secret halftone image in blockchain, a secure separable reversible data hiding (RDH) of halftone image in blockchain (SSRDHB) was proposed. A secret halftone image can be used as the original image to generate multiple share images which can be distributed storage in each point of blockchain, and additional data can be hidden to achieve management of each share image. Firstly, the secret halftone image was encrypted through Zu Chongzhi (ZUC) algorithm by using the encryption key (EK). Secondly, the method of using odd or even of share data was proposed to hide data, and a share dataset can be generated by using polynomial operation. Thirdly, multiple share images can be obtained through selecting share data, and different additional data can be hidden through controlling odd or even of share data, and additional data can be protected by using data-hiding key (DK). After sharing process, if the receiver has both keys, the halftone image can be recovered and additional data can be revealed, and two processes are separable. Experiment results show that multiple share images hidden additional data can be obtained through SSRDHB, and the halftone image can be recovered with 100% by picking any part of share images, and one additional data can be revealed with 100% by picking any one share image.
王世宇 陈浩 胡楠 贾泽坤
中国邮电高校学报(英文版), 2022, 29 (2). doi： 10. 19682/ j. cnki. 1005-8885. 2021. 0024
Due to its inherent characteristics of flexible mobility, unmanned aerial vehicle (UAV) is exploited as a cost-efficient mobile platform to assist remote data collection in the 5th generation or beyond the 5th generation (5G/ B5G) wireless systems. Compared with static terrestrial base stations, the line-of-sight (LoS) link between UAVs and ground nodes are stronger due to their flexibility in three-dimensional (3D) space. Due to the fact that flexible mobility of UAVs requires high propulsion power, the limited on-board energy constrains the performance of UAV-assisted data collection. It is worth noting that UAVs can be categorized into rotary-wing UAVs and fixed-wing UAVs, either has its own characteristics in propulsion energy consumption. In this article, a comprehensive review of state-of-art studies on trajectory design schemes for rotary-wing UAVs, as well as aerodynamic-aware attitude control strategies for fixed-wing UAVs was provided. Then, two case studies for energy-efficient data collection using rotary-wing UAVs and fixed-wing UAVs were presented, respectively. More specifically, an age-energy aware data collection scheme was demonstrated for rotary-wing UAVs to optimize the timeliness of collected data. Moreover, an aerodynamic-aware attitude control strategy for fixed-wing UAVs was also demonstrated under data collection requirements.
韩伊凡 冯涛 刘晓凯 许方敏 赵成林
中国邮电高校学报(英文版), 2022, 29 (2). doi： 10. 19682/ j. cnki. 1005-8885. 2021. 0022
With the application of various information technologies in smart manufacturing, new intelligent production mode puts forward higher demands for real-time and robustness of production scheduling. For the production scheduling problem in large-scale manufacturing environment, digital twin (DT) places high demand on data processing capability of the terminals. It requires both global prediction and rea-time response abilities. In order to solve the above problem, a DT-based edge-cloud collaborative intelligent production scheduling (DTECCS) system was proposed, and the scheduling model and method were introduced. DT-based edge-cloud collaboration (ECC) can predict the production capacity of each workshop, reassemble customer orders, optimize the allocation of global manufacturing resources in the cloud, and carry out distributed scheduling on the edge-side to improve scheduling and tasks processing efficiency. In the production process, the DTECCS system adjusts scheduling strategies in real-time, responding to changes in production conditions and order fluctuations. Finally, simulation results show the effectiveness of DTECCS system.