ID 原文 译文
17815 基于前置的攻击检测机制，获取当前ADS-B量测数据序列和预测数据序列，并在此基础上构建偏差数据序列、差分数据序列和邻近密度数据序列。依托偏差数据构建恢复向量，依托差分数据挖掘攻击数据的时序特性，依托邻近密度数据挖掘攻击数据的空间特性。 Based on attack detection strategies, the measurement andprediction sequences of ADS-B data are obtained to set up deviation data, differential data and neighbordensity data sequences, which are designed to construct recovery vectors, mine the temporal correlations andthe spatial correlations respectively.
17816 通过整合3种数据序列构建弹性恢复策略并确定恢复终止点，实现对攻击影响的弱化，将ADS-B攻击数据向正常数据方向进行定向恢复。 The selected data sequences are integrated to accomplish the wholerecovery method and decide the end point of recovery. The method is applied to eliminating attack effects and recovering the attack data towards normal data.
17817 通过对6种典型攻击样式的实验分析，证明该弹性恢复方法能够有效恢复ADS-B攻击数据，削弱数据攻击对监视系统的影响。 According to the results of experiments on six classical attackpatterns, the proposed method is effective on recovering attack data and eliminating the attack impacts.
17818 FPGA存储器映射算法负责将用户的逻辑存储需求映射到芯片中的分布式存储资源上实现。 FPGA memory mapping algorithm utilizes distributed storage resources on chip and cooperates with some auxiliary circuits to realize the different needs of users in designing logical storage functions. Previous studies on dual-port memory mapping algorithm are relatively few.
17819 前人对双端口存储器的映射算法研究相对较少，成熟的商业EDA工具的映射结果仍有不少改进空间。 There is still much space for improvementin the mapping results by mature commercial EDA tools.
17820 该文分别针对面积、延时、功耗这3个常用指标，提出一种双端口存储器映射的优化算法，并给出了具体配置方案。 An optimization algorithm of dual-port memorymapping is proposed for area, delay and power consumption, and a specific configuration scheme is given.
17821 实验表明，在面向简单存储需求时，与商用工具Vivado的映射结果一致；在面向复杂存储需求时，面积优化和功耗优化的映射结果对比商用工具改善了至少50%。 Experiments show that when facing simple storage requirements, the mapping results are consistent with thoseof commercial tools; when facing complex storage requirements, the mapping results of area optimization andpower optimization are improved by at least 50% compared with commercial tools Vivado.
17822 为提高对随机脉冲噪声(RVIN)图像的降噪效果，该文提出一种被称为双通道降噪卷积神经网络(D-DnCNN)的RVIN深度降噪模型。 A Dual-channel Denoising Convolutional Neural Network (D-DnCNN) model for the removal ofRandom-Valued Impulse Noise (RVIN) is proposed.
17823 首先，提取多个不同阶对数差值排序(ROLD)统计值及1个边缘特征统计值构成描述图块中心像素点是否为RVIN噪声的噪声感知特征矢量。 To obtain the reference image quickly, several Rank-Ordered Logarithmic absolute Difference (ROLD) statistics and one edge feature statistic are first extracted from a local window to construct a RVIN-aware feature vector which can describe the central pixel of the patchis RVIN or not.
17824 其次，利用预先训练好的深度置信网络(DBN)预测模型实现特征矢量到噪声标签的映射，完成对噪声图像中噪声点的检测。 Next, a noise detector based on Deep Belief Network (DBN) is trained to map the extractedfeature vectors to their corresponding noise labels to detect all noise-like pixels in the observed image.