ID 原文 译文
38966 以往工作中，基于组件的方法和基于注意力的方法致力于挖掘图像中的显著性区域，而忽视了用来区分易混淆类别的微弱差异。 In previous studies, the part-based and the attention-based approaches only focused on mining conspicuous regions in images, while ignoring the weak differences used to distinguish confusing categories.
38967 为了解决以上问题，本文提出了一个基于多视角融合的细粒度图像分类方法，包含两个分支，其中一个分支基于特征图挖掘图像的细粒度特征，另一个分支则学习图像的全局特征。 This paper proposed a multi-view comprehensive based fine-grained image classification model, which included two branches, one of which based on feature maps to mine fine-grained features of the image, and the other branch learned the global features of the image.
38968 同时引入一种嵌入损失，与传统多分类交叉熵损失函数结合增强特征的判别性，进而提升模型的分类性能。 A combination of embedding loss and softmax loss is introduced to enhance the discriminativeness of features, thereby improving the classification performance of the model.
38969 所提方法仅使用图像级标签，在CUB-200-2011，Stanford Cars和FGVC Aircraft这三个基准数据集上的分类准确率分别达到了88.3%，94.3%和92.4%，实验结果表明所提方法相比其他细粒度图像分类方法具有一定的优越性。 The proposed method only used image-level labels, and the classification accuracies on the three benchmarks of CUB-200-2011, Stanford Cars, and FGVC Aircraft reached 88.3%, 94.3%, and 92.4% respectively. Experimental results show that it has certain advantages for fine-grained image classification task.
38970 多功能雷达在复杂程序调度下，发射信号参数呈现取值范围宽、捷变速度快、变化随机性强等特点，非合作接收方难以对其建立有效的信号模型，给电子侦察系统的雷达辐射源识别带来严峻挑战。 Under the complex program scheduling, the multi-function radar has the characteristics of wide value range, fast agility, and strong randomness. It is difficult for non-cooperative receivers to establish an effective model of the signal, which brings serious challenges to the radar radiation source identification of electronic reconnaissance systems.
38971 本文提出一种基于深度学习的复杂体制雷达辐射源识别方法，利用大样本全脉冲数据形成脉间参数变化的图像特征表示，从宏观上揭示雷达辐射源隐含的波形设计机理，并设计了基于AlexNet网络的图像特征深度学习网络开展辐射源识别， This paper proposed a complex system radar emitter identification method based on deep learning, which used full pulse data of a large sample to form an image feature representation of pulse-to-pulse parameter changes, macroscopically revealed the waveform design mechanism implied by the radar radiation source, and designed a deep learning network of image features based on AlexNet to carry out radiation sources identification.
38972 实测数据实验表明了本文的方法对一定时间跨度内的有限部同型多功能雷达具有良好的识别性能，为多功能雷达辐射源智能个体识别提供了新的解决思路。 The measured data experiments show that the algorithm has good recognition performance for several multifunctional radars of the same type within a certain time span, which provides a new solution to the intelligent identification of multi-function radar emitter.
38973 针对现有定位算法非线性运算多，误差易传递，且存在缺秩矩阵，致使算法稳健性低、定位精度差的问题，本文提出一种站址误差条件下基于误差校正的时差定位算法。 Considering the problem that the existing localization methods have many nonlinear operations, which makes the error easy to transfer, and have rank-deficient matrix, which leads to weak robustness and low location accuracy, this paper proposes a time difference of arrival(TDOA) localization method based on error correction under the condition of sensor location error.
38974 首先通过引入辅助变量伪线性化时差定位方程，并利用加权最小二乘初步估计目标位置; Firstly, the location equation was pseudo-linearized by introducing nuisance variable, and the weighted least squares(WLS) method was used to obtain the initial estimate of the target location.
38975 接着利用辅助变量与目标位置之间的非线性函数关系，构建并求解关于目标位置误差的定位方程，对目标位置初估值进行线性校正，最终得到更为精确的估计结果。 Then, the nonlinear function relation between the nuisance variable and the target location was used to construct and solve the equation about the target location error, and the initial estimated target position was rectified linearly to obtain a more accurate final estimation result.