ID	原文	译文
53597	首先，基于稀疏特征重用、混合特征融合、中心-高斯池化三个创新点，给出了 SFＲNet 的网络结构。	First, based on the three innovations of sparse feature reuse, hybrid feature fusion, and Center-Gaussian pooling, the network structure of SFＲNet is given.
53598	然后，在图像分类数据集 ImageNet 和人脸识别数据集 LFW ( Labeled Faces in the Wild) 、MegaFace 上进行实验，分别验证了 SFＲNet 在一般场景和人脸识别这一特定场景下的特征提取能力。	Then, experiments were performed on the image classification dataset ImageNet, the face recognition dataset LFW ( Labeled Faces in the Wild) , and MegaFace, respectively, to verify the feature extraction capabilities of SFＲNet in the general scene and the specific scene of face recognition.
53599	实验表明本文所设计的 SFＲNet 不仅计算量和参数量小，还能有效提取到人脸特征并且在一般场景中也有较强的泛化能力。	Experiments show that the SFＲNet designed in this article not only has a small amount of calculation and parameters, but also can effectively extract facial features and has strong generalization ability in general scenes.
53600	相推测速技术可以实现相位量级的测量精度，在微动测量和目标识别领域有着极大的应用前景。	Phase-derived velocity measurement ( PDVM) can achieve phase level measurement accuracy, thus has a great application prospect in the field of micromotion feature extraction and target recognition.
53601	该方法对信噪比要求较高，且存在准确提取相位和解相位模糊两大难点。	The PDVM method requires high signal-to-noise ratio ( SNＲ) , and the keys of PDVM method are accurate phase extraction and resolving phase ambiguity.
53602	本文提出了针对低信噪比条件下的相推测速实现方法。	In this paper, a PDVM method based on range profiles cross correlation ( ＲPCC) under low SNＲ condition is proposed.
53603	首先，建立了宽带线性调频信号去斜处理的回波模型;	Firstly, a wideband linear frequency modulation ( LFM) signal echo model of dechirp processing is established.
53604	然后推导了相邻帧距离像互相关结果，并分析了距离像互相关输出的峰值点相位;	Then, the ＲPCC results of adjacent frames are derived, and the peak position phases of the ＲPCC results are analyzed.
53605	进而为了提高相推测速在低信噪比条件下的适用性，提出了对距离像互相关结果沿慢时间维进行积累的方法，	After that, in order to improve the applicability of PDVM under low SNＲ conditions, a method is proposed to accumulate the ＲPCC re- sults along the slow time dimension.
53606	该方法可以重新提取峰值点相位，以及获得目标速度的粗估计值进而辅助后续的解相位模糊。	This method can re-extract the peak position phases and obtain the coarse estimate of target velocity, which enables to assist in the subsequent phase ambiguity resolving processing.