Blind omnidirectional image quality assessment with representative features and viewport oriented statistical features

With the development of information technologies, various types of streaming images are generated, such as videos, graphics, Virtual Reality (VR)/omnidirectional images (OIs), etc. Among them, the OIs usually have a broader view and a higher resolution, which provides human an immersive visual experience in a head-mounted display. However, the current image quality assessment works cannot achieve good performance without considering representative human visual features and visual viewing characteristics of OIs, which limited OIs’ further development. Motivated by the above problem, this work proposes a blind omnidirectional image quality assessment (BOIQA) model based on representative features and viewport oriented statistical features. Specifically, we apply the local binary pattern operator to encoder the cross-channel color information, and apply the weighted LBP to extract the structural features. Then the local natural scene statistics (NSS) features are extracted by using the viewport sampling to boost the performance. Finally, we apply support vector regression to predict the OIs’ quality score, and experimental results on CVIQD2018 and OIQA2018 Databases prove that the proposed model achieves better performance than state-of-the-art OIQA models.     

Ultra-Wideband Electrostrictive Mechanical Antenna

Mechanical antennas based on piezoelectric materials can effectively reduce the size of long wave antennas down to 1/1000 of the wavelength (from km scale to mm level). However, the narrow bandwidth and weak field intensity seriously restrict its practical applications in transmission distance and channel capacity. Here, a mechanical antenna-based electrostrictive effect of relaxor ferroelectric ceramic (PMN-PT) is proposed to improve radiation capacity and achieve ultra-wideband characteristics (10 kHz–1 MHz). Due to the ultra-high dielectric constant at working temperature and the relationship between the strain and applied field intensity, the proposed antenna gets rid of the dependence on the poled materials and exhibits excellent communication properties beyond traditional mechanical antennas, which are experimentally demonstrated by a practical wireless communication system. Only using a single proposed mechanical antenna with 8 mm diameter and 3 mm thickness, the effective communication with a transmission distance of 200 m can be realized. This design offers a promising way of constructing mechanical antennas for long-wave communication.     

Toward Fine-grained Image Retrieval with Adaptive Deep Learning for Cultural Heritage Image

Fine-grained image classification is a challenging research topic because of the high degree of similarity among categories and the high degree of dissimilarity for a specific category caused by different poses and scales. A cultural heritage image is one of the fine-grained images because each image has the same similarity in most cases. Using the classification technique, distinguishing cultural heritage architecture may be difficult. This study proposes a cultural heritage content retrieval method using adaptive deep learning for fine-grained image retrieval. The key contribution of this research was the creation of a retrieval model that could handle incremental streams of new categories while maintaining its past performance in old categories and not losing the old categorization of a cultural heritage image. The goal of the proposed method is to perform a retrieval task for classes. Incremental learning for new classes was conducted to reduce the re-training process. In this step, the original class is not necessary for re-training which we call an adaptive deep learning technique. Cultural heritage in the case of Thai archaeological site architecture was retrieved through machine learning and image processing. We analyze the experimental results of incremental learning for fine-grained images with images of Thai archaeological site architecture from world heritage provinces in Thailand, which have a similar architecture. Using a fine-grained image retrieval technique for this group of cultural heritage images in a database can solve the problem of a high degree of similarity among categories and a high degree of dissimilarity for a specific category. The proposed method for retrieving the correct image from a database can deliver an average accuracy of 85 percent. Adaptive deep learning for fine-grained image retrieval was used to retrieve cultural heritage content, and it outperformed state-of-the-art methods in fine-grained image retrieval.     

太赫兹电波传播及信道特性

首先,综述太赫兹电波相较于低频电波传播的不同特性,包括气象因素和材料粗糙表面对电磁波的影响。接着,提出利用射线跟踪技术仅通过有限的信道测量数据校准目标场景中的三维环境模型以及材料电磁参数;然后,利用从射线跟踪仿真反演出的参数在类似场景中进行大量仿真,代替信道测量生成大量真实有效的全维度信道数据;最后,提取并分析信道特性,例如路径损耗、阴影衰落、莱斯K因子、均方根时延扩展、角度扩展及多普勒参数。2个案例研究是从室内桌面通信到室外智能车联网场景,分别代表了6G移动通信从近到远用例的两端,对于室外场景还额外考虑了不同气象条件下对信道参数的影响,对太赫兹系统的设计和评估具有重大意义。     

基于自适应方向导数滤波器的彩色边缘检测

传统彩色边缘检测算法在提高边缘检测准确性时可能将噪声检测为边缘,而在提高噪声鲁棒性时会将部分边缘当作噪声进行抑制,导致部分边缘信息丢失。为解决传统彩色边缘检测算法在边缘检测准确性与噪声鲁棒性之间的矛盾问题,提出一种基于自适应各向异性高斯方向导数（ANDD）的彩色边缘检测算法。通过彩色图像的微分自相关矩阵构建反映边缘类型的度量准则,以自适应地确定每个像素处ANDD滤波器的形状,从而准确提取不同类型的边缘特征,采用ANDD滤波器组对图像进行平滑处理,提取在三个通道上的ANDD特征。在此基础上,利用奇异值分解得到最优融合权值,并融合三个通道的ANDD特征,以增强彩色边缘强度。实验结果表明,该算法在无噪声和含噪声环境下的Pratt品质因子分别为0.849 6和0.791 4,与彩色Canny、RCMG-MM和FRPOS算法相比,在保持较高边缘检测准确率的同时具有较优的噪声鲁棒性。     

Factors contributing to young moped rider accidents in Denmark

Young road users still constitute a high-risk group with regard to road traffic accidents. The crash rate of a moped is four times greater than that of a motorcycle, and the likelihood of being injured in a road traffic accident is 10–20 times higher among moped riders compared to car drivers. Nevertheless, research on the behaviour and accident involvement of young moped riders remains sparse.Based on analysis of 128 accident protocols, the purpose of this study was to increase knowledge about moped accidents. The study was performed in Denmark involving riders aged 16 or 17. A distinction was made between accident factors related to (1) the road and its surroundings, (2) the vehicle, and (3) the reported behaviour and condition of the road user. Thirteen accident factors were identified with the majority concerning the reported behaviour and condition of the road user. The average number of accident factors assigned per accident was 2.7. Riding speed was assigned in 45% of the accidents which made it the most frequently assigned factor on the part of the moped rider followed by attention errors (42%), a tuned up moped (29%) and position on the road (14%). For the other parties involved, attention error (52%) was the most frequently assigned accident factor. The majority (78%) of the accidents involved road rule breaching on the part of the moped rider.The results indicate that preventive measures should aim to eliminate violations and increase anticipatory skills among moped riders and awareness of mopeds among other road users. Due to their young age the effect of such measures could be enhanced by infrastructural measures facilitating safe interaction between mopeds and other road users.     

基于 ResU-Net 的三维断层识别方法及应用

断层刻画了地层的边界位置,地震成像数据中反射层的不连续性可作为断层解释的主要依据。深度神经网络的强非线性性质可作为地震数据中断层不连续特征表达的有力工具,断层识别问题可视作一个像素级别的二分类问题,并使用深度学习方法对此问题进行建模求解。据此可给出一种端到端的基于深度学习网络的三维断层自动识别方法。首先利用地震子波与反射系数卷积合成多组三维地震数据,建立用于深度网络学习断层特征的样本数据,随后搭建网络进行训练,网络训练完成后应用于实际地震数据。鉴于残差模块可很好地提升网络泛化性能,所提出的将残差网络中的残差块结构引入U-Net中的方法,可用于提升通过合成数据样本训练得到的网络模型在训练数据之外,即实际地震数据上的断层识别性能。所建立网络用于断层解释时,输入为叠后三维地震数据,输出为相同维度的三维数据体,其中每一输出值代表输入三维地震数据相同位置处断层的概率。实际算例对比测试表明,此方法可对三维地震数据中的断层进行有效识别,在合成数据集上训练精度相差不大的前提下,引入残差模块的ResU-Net在实际地震数据上的断层识别泛化性能得到提升。     

A soft computing approach for privacy requirements engineering: The PriS framework

Soft computing continuously gains interest in many fields of academic and industrial domain; among the most notable characteristics for using soft computing methodological tools is the ability to handle with vague and imprecise data in decision making processes. Similar conditions are often encountered in requirements engineering. In this paper, we introduce the PriS approach, a security and privacy requirements engineering framework which aims at incorporating privacy requirements early in the system development process. Specifically, PriS provides a set of concepts for modelling privacy requirements in the organisation domain and a systematic way-of-working for translating these requirements into system models. The conceptual model of PriS uses a goal hierarchy structure. Every privacy requirement is either applied or not on every goal. To this end every privacy requirement is a variable that can take two values [0,1] on every goal meaning that the requirements constraints the goal (value 1) or not (value 0). Following this way of working PriS ends up suggesting a number of implementation techniques based on the privacy requirements constraining the respective goals. Taking into account that the mapping of privacy variables to a crisp set consisting of two values [0,1] is constraining, we extend also the PriS framework so as to be able to address the degree of participation of every privacy requirement towards achieving the generic goal of privacy. Therefore, we propose a fuzzification of privacy variables that maps the expression of the degree of participation of each privacy variable to the [0,1] interval. We also present a mathematical framework that allows the concurrent management of combined independent preferences towards the necessity of a privacy measure; among the advantages of the presented extended framework is the scalability of the approach in such a way that the results are not limited by the number of independent opinions or by the number of factors considered while reasoning for a specific selection of privacy measures.     

采用独立分量分析Zernike矩的遥感图像飞机目标识别

为了提高遥感图像目标自动识别系统的准确性,提出了一种新的采用独立分量分析(ICA)Zern ike矩的飞机目标识别方法.首先对分割后的目标区域进行独立分量分析处理,将待识别目标形状转换到标准形式,然后对标准化后的图像目标提取Zern ike矩作为特征向量进行识别.通过实验表明此方法具有鲁棒性,能有效地消除遥感图像目标尺度、旋转、平移、反转和扭曲影响,能够有效地识别遥感图像飞机目标.