ID 原文 译文
2723 目前，人脸美丽预测存在数据规模小、分类难度大、深度特征研究不足等问题。 At present, facial beauty prediction is facing the problems, in which data is insufficient, the face image is hard to classify, and the deep feature lacks research.
2724 为此，本文提出基于双激活层深度卷积特征的人脸美丽预测研究的解决方案。 To solve these problems, a solution to facial beauty prediction research based on double activation layer depth convolution feature is proposed.
2725 首先，采用数据增强和人脸对齐方法来增加训练集的样本数量和提高数据库的数据质量。 Firstly, we use the method of data augmentation andface alignment to increase the number of samples in training set and improve the data quality of database.
2726 其次，提出一种双激活层改进 CNN 模型，使其更适合人脸美丽预测应用。 Secondly, we pro-pose a double activation layer (DAL)to design a CNN model that is more suitable for facial beauty prediction.
2727 实验结果表明，本文所提方法在分类和回归预测方面均大幅度优于传统人脸美丽预测方法; Experimen-tal results based on 2000 test set show that the method proposed is superior to the traditional method of facial beauty predic-tion both in classification and regression.
2728 同时，在主流的 CNN 模型中取得了较好的实时性和准确性，基于 2000 测试集的分类准确率达到 61.1% ，回归相关度达到 0.8546。 In addition, the proposed method achieves better results and real time performance than the state-of-art CNN model, in which rank-1 recognition rate is 61. 1% and the Pearson correlation coefficient is 0. 8546.
2729 因此，双激活层在深层人脸美丽特征学习中发挥了重要作用，可广泛应用于人脸图像识别与处理。 Consequently, the DAL method plays an important role in deep facial prediction learning, which can be widely used in face recognition and image processing.
2730 针对谱峰搜索的二维波达方向估计中现有算法复杂度高，精度受搜索间隔影响较大的问题，给出了一种双向传播算子的互质面阵二维波达方向估计算法，实现了俯仰角和方位角的低复杂、高精度、无模糊联合估计。 In the existing two-dimensional direction-of-arrival (DOA)estimation algorithms based on spectral peaksearch, the complexity is high and the accuracy is greatly influenced by the search interval. To overcome these problems, This paper examines a two-dimensional DOA estimation with coprime rectangular array using bi-directional propagator method, which realizes a low complexity, high accuracy, unambiguous for 2-D estimation.
2731 该方法首先将互质阵列引入到二维波达方向估计中，构造互质平面阵模型， Firstly, this algorithm introduces coprimearray into 2-D DOA estimation, and constructs a coprime rectangular array model.
2732 然后采用两次旋转不变传播算子方法计算出不同阵列流型方向上的旋转因子矩阵，根据旋转因子矩阵解算出目标信号的俯仰角和方位角， The two rotation factor matrices along the different directions for propagator method can be got, the elevation angle and azimuth angle for the sources can be obtained from the rotation factor matrices.