ID 原文 译文
17825 再次，在噪声检测标签的指示下采用Delaunay三角剖分插值算法快速修复噪声像素点从而获得初步复原图像。 Then, under the guidance of noise labels, the Delaunay triangulation-based interpolation algorithm is exploited to restore all detected noise-like pixels quickly and generate a preliminary restored image used as reference image.
17826 最后，将初步复原图像作为参考图像与噪声图像联接(concatenate)后输入D-DnCNN模型后获得残差图像，将参考图像减去残差图像即可获得降噪后图像。 Finally, the reference image and the noisy image are simultaneously fed into the D-DnCNN model to output itscorresponding residual image, and the final restored image can be obtained by subtracting the residual imagefrom the noisy image.
17827 实验数据表明：D-DnCNN模型在各个噪声比例下的降噪效果均显著超过了现有的经典开关型RVIN降噪算法，与普通的单通道RVIN深度降噪模型相比也有较大幅度提升。 Extensive experimental results show that, the denoising effect of the proposed D-DnCNNdenoising model outperforms the existing state-of-art switching ones across a range of noise ratios, and it alsoworks better than the ordinary single-channel DnCNN model.
17828 移动云计算、人工智能(AI)、5G等新兴技术应用促使弹性光网络(EON)在骨干传输网中发挥更重要的角色，降级服务(DS)技术为降低EON的业务阻塞率、提高频谱利用率提供了新途径。 Emerging technology applications such as mobile cloud computing, Artificial Intelligence (AI) and 5Gpromot Elastic Optical Network (EON) to play an important role in the backbone transmission network.Degraded Service (DS) technology can provide a new way to reduce traffic congestion and improve spectrumutilization in EON.
17829 该文首先对现有DS算法的资源分配不公、忽略低等级业务的体验质量(QoE)等问题，建立了以最小化降级频次、降级等级与传输时延损失(TDL)为联合优化目标的混合整数线性规划(MILP)模型，并提出一种时延感知的降级恢复路由与频谱分配(DDR-RSA)算法。 Firstly, considering the problems of unfair resource allocation and neglecting the Quality ofExperience (QoE) of low-priority services in existing DS algorithms, a Mixed Integer Linear Program (MILP)model with the joint objective of minimizing downgrade frequency, downgrade level and Transmission DelayLoss (TDL) is established. A Delay-aware Degradation-Recovery Routing and Spectrum Assignment (DRR-RSA) algorithm for degraded recovery is proposed.
17830 为提高降级业务的QoE和运营商收益，在算法的最优DS窗口选择阶段中融入降级恢复策略， In order to improve the QoE of downgraded services andthe revenue of operators, the strategy of degradation recovery is integrated in the optimal DS-window selectionphase of the algorithm.
17831 在保障传输数据量不变的前提下，将降级业务向空闲频域复原，从而提高频谱效率、减小降级业务TDL和最大化网络收益。 Under the premise of guaranteeing the transmission data quantity unchanged, thedegradable services are restored to the free spectrum domain, so as to increase the spectrum efficiency, reducedegraded service TDL and maximize revenue.
17832 最后，通过仿真证明了所提算法在业务阻塞率、网络收益和降级业务成功率等方面的优势。 Finally, the simulation results tesfity that the proposed algorithmhas advantages in terms of traffic congestion, revenue and degraded service success-rate.
17833 云计算作为一种新型计算模式，具有海量资源、动态扩展、按需分配等特点。 Cloud computing, as a new computing paradigm, offers dynamically scalable and seemly unboundedstorage and computation resources in a pay-as-you-go manner.
17834 资源受限的用户可以将计算任务外包给云服务器，在享受高质量数据服务的同时大大降低了本地管理开销。 In order to enjoy superior data services and reduce the local maintenance cost, more and more resource-constrained users prefer to outsource their data to the cloud server.