ID 原文 译文
22455 该文针对现有穿墙雷达建筑物布局成像中扩展目标稀疏成像方法未能有效利用墙体反射信号的结构稀疏性，导致成像中出现墙体不连贯和墙体轮廓不明显的问题，提出一种利用稀疏信号块间耦合的建筑物布局成像方法。 In through-wall radar building layout imaging, the existing extended target sparse imaging method can not effectively exploit the structural sparsity of the wall reflections in the scene, resulting in incoherent imaging and unobvious contour of walls. A sparse Bayesian learning method is proposed for building layout imaging by exploiting the inter-block coupling of sparse signal.
22456 该方法在块稀疏信号特性的高斯分层先验模型的基础上，利用块间耦合系数进一步表征场景中墙体反射信号的结构稀疏性，然后将其引入到控制稀疏信号先验概率分布的超参数中，从而把稀疏信号的结构性转化为超参数的耦合关系，最后利用期望最大化(EM)算法求解超参数的最大后验(MAP)估计。 On the basis of the hierarchical Gaussian prior model of block sparse signal characteristics, the inter-block coupling coefficient is further used to characterize the structured sparsity of the wall reflections. Then these coefficients are introduced into the hyperparameters controlling the prior distribution of sparse signal, thus this structured sparsity is transformed into the coupling relationship of these hyperparameters. Susequently, an Expectation-Maximization (EM) algorithm is developed to infer the Maximum A Posterior (MAP) estimate of these hyperparameters.
22457 仿真和实验数据处理结果表明，该方法有效改善了墙体的成像质量。 The results of simulation and experiment show that the proposed method improves effectively the imaging quality of the building wall.
22458 综合考虑内容中心网络(CCN)的能耗优化及性能提升，该文提出一种内容中心网络中能耗优化的隐式协作缓存机制。 Taking  into  account  the  energy  optimization  and  performance  enhancement  of  the  Content  Centric Networking  (CCN)  comprehensively,  an  energy  optimized  implicit  collaborative  caching  scheme  for  CCN  is proposed.
22459 缓存决策时，利用缓存节能量作为判决条件优先在用户远端节点缓存，并利用数据包携带最近上游缓存跳数信息进行隐式协作，减轻用户近端节点缓存空间的竞争压力，提高邻近节点缓存的差异性。 In terms of the caching decision, energy saving account is utilized as the judgement, which is carried out on consumer's remote nodes preferentially, and the data packet is utilized to carry the information of recent upstream caching hops, so as to realize the implicit collaboration, thus reducing the caching space competition pressure of the consumers' near nodes, improving the caching difference between nearby nodes.
22460 缓存替换时，选取缓存节能量最小的缓存内容加以替换，达到最优的能耗优化效果。 As for the caching replacement, the caching content with the minimum energy saving account is selected to be replaced, achieving the optimal energy consumption optimization effect.
22461 仿真结果表明，该缓存机制在性能上获得较优的缓存命中率及平均路由跳数，同时有效降低了网络能耗。 Simulation results show that, the caching scheme achieves better cache hit ratio and average routing hops, meanwhile, it reduces the network energy consumption effectively.
22462 手势识别作为人机交互的方式之一，在人工智能日益发展的今天备受瞩目。 As an important milestone for the development of the artificial intelligence, gesture recognition enables the human-computer interaction and has received significantly growing research interest nowadays.
22463 针对手势旋转、平移、缩放等形变导致识别率偏低的问题，该文基于密集卷积网络(Densenet)与空间转换网络(STN)提出了一种新型的网络结构Densenet_V2，先利用空间转换网络对输入的样本和特征图进行空间变换和对齐，再利用密集卷积网络自动提取手势的有效特征，最后通过线性分类器对手势进行分类。 However, the current technology for the gesture recognition has the low quality in the gesture rotation, translation and scaling. To solve the problem, a novel network structure named Densenet_V2 is proposed, and it is based on Dense Convolutional Networks (Densenet) and Spatial Transformer Networks (STN). Firstly, the input samples and feature maps are spatially transformed and aligned with the STN. Then the effective features of gestures are automatically extracted by using the Densenet. Finally, the linear classifier is adopted to classify the gestures.
22464 为防止网络模型对样本数据集过度拟合，对网络进行训练时在损失函数中加入 L2 正则项以实现权重衰减。在 Marcel 手势库上进行多次实验。 To prevent the network model from over-fitting the sample data set, the L2 regular term is involved into the loss function to achieve the weight decay when training the network.