摘 要：文章对基于深度学习的亲属关系验证方法进行了深入研究，并针对由于人脸图像与其他自然图像存在较大的差异而导致的感受野较小的问题，提出了一种基于空洞卷积神经网络的亲属关系验证方法，构建了残差空洞卷积神经网络（RDCN Net），分别从父母与孩子的人脸图像中提取深度特征，经过特征融合后使用鉴别器得到亲属关系验证结果。算法在公开亲属关系数据集 KinFaceW 上进行测试，实验结果表明，本文方法在亲属关系验证的准确率上有良好的表现。
关键词：亲属关系验证；深度特征；空洞卷积；特征融合
DOI:10.19850/j.cnki.2096-4706.2021.18.019
基金项目：河南省科技攻关项目（1921022 10277）；河南省自然科学基金面上项目（20230 0410092）
中图分类号：TP391 文献标识码：A 文章编号：2096-4706（2021）18-0071-05
Kinship Verification Method Based on Atrous Convolution Neural Network
ZHENG Liang1 , CHEN Peng2 , HAN Jingjing3 , CHEN Ya 1
(1.School of Computer and Information Engineering, Henan University, Kaifeng 475001, China; 2.Polytechnic College, Hebei University of Science and Technology, Shijiazhuang 050018, China; 3. Xinxian Middle School, Xinyang 465550, China)
Abstract: In this paper, the kinship verification method based on deep learning is deeply studied, and aiming at the problem of small receptive field caused by the large difference between face image and other natural images, a kinship verification method based on atrous convolution neural network is proposed, residual atrous convolutions neural network (RDCN Net) is constructed, the depth features are extracted from the face images of parents and children respectively, and the kinship verification results are obtained by using the discriminator after feature fusion. The algorithm is tested on the open kinship dataset KinFaceW. The experimental results show that the proposed method has good performance in the accuracy of kinship verification.
Keywords: kinship verification; deep feature; atrous convolution; feature fusion
参考文献：
[1] MARTELLO M F D，MALONEY L T. Where are kin recognition signals in the human face? [J/OL].Journal of Vision，2006， 6（12）：1356-1366.[2021-06-22].http://journalofvision.org/6/12/2/.
[2] MARTELLO M F D，MALONEY L T. Lateralization of kin recognition signals in the human face [J/OL].Journal of vision，2010，10（8）：1-10.[2021-06-22].http://www.journalofvision.org/ content/10/8/9.
[3] DEBRUINE L M，SMITH F G，Jones B C，et al. Kin recognition signals in adult faces [J].Vision Research，2009，49（1）： 38-43.
[4] MALONEY L T，MARTELLO M F D. Kin recognition and the perceived facial similarity of children [J/OL].Journal of Vision， 2006，6（10）：1047-1056.[2021-06-22].http://journalofvision. org/6/10/4/.
[5] ALVERGNE A，PERREAU F，MAZUR A，e t al.Identification of visual paternity cues in humans [J/OL].Biology letters，2014，10（4）：1-4.[2021-06-22].https://doi.org/10.1098/ rsbl.2014.0063.
[6] HE K M，ZHANG Y，REN S Q，et al. Deep Residual Learning for Image Recognition [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition（CVPR）.Las Vegas： IEEE，2016：770-778.
[7] FANG R，TANG K D，SNAVELY N，et al. Towards computational models of kinship verification [C]//2010 IEEE International Conference on Image Processing.Hong Kong：IEEE， 2010：1577-1580.
[8] CHEN X P，ZHU X K，ZHENG S S，et al. Semi-Coupled Synthesis and Analysis Dictionary Pair Learning for Kinship Verification [J].IEEE Transactions on Circuits and Systems for Video Technology， 2021，31（5）：1939-1952.
[9] LU J W，LIONG V E，ZHOU X Z，et al. Learning Compact Binary Face Descriptor for Face Recognition [J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2015，37（10）：2041- 2056.
[10] ZHANG K H，HUANG Y Z，SONG C F，et al. Kinship Verification with Deep Convolutional Neural Networks [C]//Proceedings of the British Machine Vision Conference（BMVC）Swansea， Swansea：BMVA Press，2015：148.1-148.12.
[11] ZHOU X Z，JIN K，XU M，et al. Learning Deep Compact Similarity Metric for Kinship Verification from Face Images [J]. Information Fusion，2019，48：84-94.
[12] YU J，LI M Y，HAO X L，et al. Deep Fusion Siamese Network for Automatic Kinship Verification [J/OL].arXiv：2006.00143 [cs.CV].[2021.06-22].https://arxiv.org/abs/2006.00143v2.
[13] LU J W，ZHOU X Z，TAN Y P，et al. Neighborhood Repulsed Metric Learning for Kinship Verification [J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2014，36（2）：331- 345.
[14] ZHOU X Z，YAN H B，SHANG Y Y. Kinship verification from facial images by scalable similarity fusion [J].Neurocomputing， 2016，197：136-142.
[15] NANDY A，MONDAL S S. Kinship Verification using Deep Siamese Convolutional Neural Network [C]//International Conference on Automatic Face and Gesture Recognition.Lille：[s.n.]，2019：1-5.
[16] BENGIO Y，SIMARD P，FRASCONI P. Learning long-term dependencies with gradient descent is difficult [J].IEEE Transactions on Neural Networks，1994，5（2）：157-166.
[17] GLOROT X， BENGIO Y. Understanding the difficulty of training deep feedforward neural networks [J].Journal of Machine Learning Research，2010，9：249-256.
[18] LECUN Y，BOSER B E，DENKER J S，et al. Backpropagation applied to handwritten zip code recognition [J].Neural computation，1989，1（4）：541-551.
[19] YU F，KOLTUN V，FUNKHOUSER T. Dilated Residual Networks [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition（CVPR），Honolulu：IEEE，2017：636-644.
[20] HINTON G E，SRIVASTAVA N，KRIZHEVSKY A，et al.Improving neural networks by preventing co-adaptation of feature detectors [J/OL].arXiv：1207.0580 [cs.NE].[2021-06-22].https://arxiv. org/abs/1207.0580.
[21] DEHGHAN A，ORTIZ E G，VILLEGAS R，et al. Who Do I Look Like? Determining Parent-Offspring Resemblance via Gated Autoencoders [C]//2014 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）.Columbus：IEEE，2014：1757- 1764.
[22] YAN H B，LU J W，DENG W H，et al. Discriminative Multimetric Learning for Kinship Verification [J].IEEE Transactions on Information Forensics and Security，2014，9（7）：1169-1178.
[23] YAN H B，LU J W，ZHOU X Z. Prototype-Based Discriminative Feature Learning for Kinship Verification [J].IEEE Transactions on Cybernetics，2015，45（11）：2535-2545.
作者简介：郑亮（1993 －），男，汉族，河南信阳人，硕士研究生在读，研究方向：模式识别，深度学习；陈鹏（1999 －），男， 汉族，河北石家庄人，本科在读，研究方向：网络安全，深度学习； 韩晶晶（1989 －），女，汉族，河南信阳人，教师，本科，研究方 向：教育信息化；陈亚（1993 －），女，汉族，河北石家庄人，硕士研究生在读，研究方向：模式识别，深度学习。