ID	原文	译文
3833	针对核范数有偏近似秩函数导致基于核范数最小化的传统去噪方法去噪性能较差的问题，基于低秩理论，提出一种基于伽马范数最小化的图像去噪算法。	Focusing on the issue of rather poor denoising performance of the traditional kernel norm minimization basedmethod caused by the biased approximation of kernel norm to rank function, based on the low-rank theory, a gammanorm minimization based image denoising algorithm was developed.
3834	首先对噪声图像重叠分块，然后基于结构相似性指数自适应搜索与当前图像块最相似的若干非局部图像块以组成相似图像块矩阵，	The noisy image was firstly divided into some overlapping patches via the proposed algorithm, and then several non-local image patches most similar to the current im-age patch were sought adaptively based on the structural similarity index to form the similar image patch matrix.
3835	进而利用非凸伽马范数无偏近似矩阵秩函数构建低秩去噪模型，	Subse-quently, the non-convex gamma norm could be exploited to obtain unbiased approximation of the matrix rank functionsuch that the low-rank denoising model could be constructed.
3836	最后基于奇异值分解对所得低秩去噪优化问题求解，并将去噪图像块重组为去噪图像。	Finally, the obtained low-rank denoising optimization issuecould be tackled on the basis of singular value decomposition, and therefore the denoised image patches could bere-constructed as a denoised image.
3837	仿真结果表明，与现有主流 PID、NLM、BM3D、NNM、WNNM、DnCNN 和 FFDNet 算法相比，所提算法可较显著地消除高斯噪声，且可较好地恢复原始图像细节。	Simulation results demonstrate that, compared to the existing state-of-the-art PID,NLM, BM3D, NNM, WNNM, DnCNN and FFDNet algorithms, the developed method can eliminate Gaussian noisemore considerably and retrieve the original image details rather precisely.
3838	针对动态广域光骨干网中光信道传输质量预测方法精确度不足的问题，以集成学习理论为基础提出一种光信道传输质量预测方法。	Due to the fact that in dynamic wide-area optical backbone network the accuracies of the existing predictionmethods were insufficient, a novel prediction method on quality of transmission (QoT) of optical channel was proposedbased on ensemble learning theory.
3839	首先，在堆栈集成学习框架下构建了由 5 个多层感知器模型组成的基学习器，通过并行组合的方式实现了样本数据的同态集成学习。	Firstly, under the framework of stacked ensemble learning, a base-learner including fivemultilayer perceptron (MLP) model was built, which could achieve homomorphic ensemble learning of sample data throughparallel combination.
3840	然后，融合基学习器的预测结果形成新的训练集，用于训练由单一多层感知器组成的元学习器。	Subsequently, the new training set fused from the predicted results of the preceding base-learner wasused to training the meta-learner composed of a single MLP.
3841	仿真结果表明，对比深度神经网络，所提方法在单信道和多信道 QoT 预测场景下具有更优秀的非线性逼近性能，预测精度分别提高了 1.93%和 3.82%。	The simulation results show that compared with the used deepneural network, the proposed method can obtain a more excellent nonlinear approximation in the scenarios of the sin-gle-channel and multi-channels, and the prediction accuracies have the improvements of 1.93% and 3.82% respectively.
3842	针对应急物联网（EIoT）超低时延服务需求，设计了面向超低时延传输应急物联网的多切片网络架构，提出 EIoT 切片资源预留和多异构切片资源共享与隔离的方法框架。	Based on the requirements of ultra-low latency services for emergency Internet-of-things (EIoT) applica-tions, a multi-slice network architecture for ultra-low latency emergency IoT was designed, and a general methodologyframework based on resource reservation, sharing and isolation for multiple slices was proposed.