计算机科学 ›› 2020, Vol. 47 ›› Issue (9): 163-168.doi: 10.11896/jsjkx.190900118

• 计算机图形学&多媒体 • 上一篇    下一篇

基于深度学习与多哈希相似度加权实现快速人脸识别

邓良1, 许庚林1, 李梦杰1, 陈章进1,2   

  1. 1 上海大学微电子研究与开发中心 上海200444
    2 上海大学计算机中心 上海200444
  • 收稿日期:2019-09-18 发布日期:2020-09-10
  • 通讯作者: 陈章进(zjchen@shu.edu.cn)
  • 作者简介:1593481663@qq.com
  • 基金资助:
    国家自然科学基金(61674100)

Fast Face Recognition Based on Deep Learning and Multiple Hash Similarity Weighting

DENG Liang1, XU Geng-lin1, LI Meng-jie1, CHEN Zhang-jin1,2   

  1. 1 Microelectronics Research and Development Center,Shanghai University,Shanghai 200444,China
    2 Computer Center,Shanghai University,Shanghai 200444,China
  • Received:2019-09-18 Published:2020-09-10
  • About author:DENG Liang,born in 1996,master.His main research interests include digital chip design,deep learning,and face recognition.
    CHEN Zhang-jin,born in 1969,doctor,professor.His main research interests include digital chip design,large-screen LED display research and development.
  • Supported by:
    National Natural Science Foundation of China (61674100).

摘要: 无论是使用传统的方法进行人脸识别,还是使用神经网络进行人脸识别,都存在运算量大、运算时间长等问题,很难对视频中的人脸进行实时检测与匹配。针对上述问题,使用轻量化神经网络进行人脸检测,使用运算简单的哈希算法计算人脸图像相似度,并对多个哈希相似度值加权进行人脸匹配,是减少运算时间、实现快速人脸识别的可行方案。使用轻量化神经网络Mobilenet作为人脸特征提取网络,使用剪枝的SSD模型作为检测网络,通过级联Mobilenet与SSD实现人脸的检测,之后对检测到的人脸图像进行识别。首先,分别计算人脸图像的均值哈希相似度与感知哈希相似度。然后,分别使用αβ作为均值哈希与感知哈希的加权系数对图像的均值哈希与感知哈希相似度值进行加权,并将结果作为图像的最终相似度。当加权后的相似度值大于设定的阈值I时,则认为两张图像中的人脸是同一个人;当加权后的相似度值小于设定的阈值K时,则认为两张图像中的人脸是不同的人。对于相似度处于阈值I和阈值K之间的图像,将它们按照相似度值从高到低的顺序择优匹配。所提方法在WiderFace和FDDB上的人脸检测准确率分别达到92.5%和94.2%,每张图片的平均处理时间为56ms;在ORL标准人脸库进行人脸匹配的准确率达到96.2%。使用摄像头进行实时人脸识别测试时,所提方法的人脸识别准确率为95%,平均人脸识别速度为80ms。实验证明,所提方法在保证较高准确率的前提下,能够实现实时的人脸检测与匹配。

关键词: 人脸检测, 深度学习, 哈希算法, 人脸匹配

Abstract: Whether using the traditional method or neural network for face recognition,there are problems of large computation and long computation time.It is difficult to detect and match the faces in the video in real time.Aiming at the above problems,lightweight neural network is used for face detection,simple hash algorithm is used to calculate the similarity of face images,and multiple hash similarity values are weighted for face matching.It is a feasible scheme to reduce computation time and realize fast face recognition.The lightweight neural network Mobilenet is used as the face feature extraction network,and the pruned SSD model is used as the detection network.The face detection is realized by cascading Mobilenet and SSD,and then the detected face image is recognized.Firstly,the mean hash similarity and the perceived hash similarity of the face images are calculated separately.Then,taking α and β as weighted coefficients of the mean hash and the perceived hash respectively,the mean hash and perceived hash similarity value of the image are weighted,and the result is taken as the final similarity of the image.When the weighted similarity value is greater than the set threshold I,it is considered to be the same person.When the weighted similarity value is less than the set threshold K,it is considered to be a different person.For images whose similarity is between thresholds I and K,they are optimally matched in order of similarity values from high to low.The face detection accuracy rate of the proposed method on WiderFace and FDDB reaches 92.5% and 94.2% respectively,and the average processing time per image is 56ms.The accuracy of face matching in the ORL standard face database reaches 96.2%.When camera is used for real-time face recognition test,the face recognition accuracy of the proposed method is 95%,and the average face recognition speed is 80ms.It has been proved by experiments that real-time face detection and matching can be realized under the premise of ensuring high accuracy.

Key words: Face detection, Deep learning, Hash algorithm, Face matching

中图分类号: 

  • TP183
[1] JIANG R D,CHU W,YAN J.Intelligent recognition system for face recognition under ARM platform[J].Computer Engineering and Design,2018,39(2):590-595.
[2] BUHMANN J,LADES M,MALSBURG C V D.Size and distortion invariant object recognition by hierarchical graph matching[C]//International Joint Conference on Neural Networks.1990:416-441.
[3] SONG J,CHEN B,CHI Z,et al.Face Recognition Based on Binary Template Matching[C]//Third International Conference on Intelligent Computing(ICIC 2007).Springer-Verlag,2007.
[4] SHAN S,GAO W,ZHAO D.Face recognition based on face-specific subspace[J].International Journal of Imaging Systems &Technology,2010,13(1):23-32.
[5] DAVE P,AGARWAL J,METTA T.Face Detection using Principal Component Analysis (PCA)[J].International Journal of Computer Applications,2014,95(17):37-40.
[6] RANGSEE P,RAJA K B,VENUGOPAL K R.Modified Local Ternary Pattern Based Face Recognition Using SVM[C]//2018 International Conference on Intelligent Informatics and Biome-dical Sciences (ICIIBMS).2018.
[7] REN S,HE K,GIRSHICK R,et al.Faster R-CNN:TowardsReal-Time Object Detection with Region Proposal Networks[C]//International Conference on Neural Information Processing Systems.2015.
[8] REDMON J,DIVVALA S,GIRSHICK R,et al.You Only Look Once:Unified,Real-Time Object Detection[J].arXiv:1506.02640,2015.
[9] LIU W,ANGUELOV D,ERHAN D,et al.SSD:Single ShotMultiBox Detector[J].arXiv:1512.02325,2015.
[10] SIMONYAN K,ZISSER,AN A.Very Deep Convolutional Networks for Large-Scale Image Recognition[J].arXiv:1409.1556v6,2014.
[11] HUANG J H,LI X W,CHEN B H,et al.Comparative Study of Image Similarity Algorithm Based on Hash[J].Journal of Dali University,2017,12(2):32-37.
[12] YIN Y M,PENG Y,QI J H.Research on the Image Similarity Retrieval Algorithm Based on Double Hash[J].Information and Communications Technologies,2019,13(1):33-38.
[13] DENG J,ZHOU B,SHI Y L.Application of Improved Image Hash Algorithm in Image Tamper Detection[C]//2018 International Conference on Intelligent Transportation,Big Data & Smart City (ICITBS).IEEE Computer Society,2018.
[14] HOWARD A G,ZHU M,CHEN B,et al.MobileNets:Efficient Convolutional Neural Networks for Mobile Vision Applications[J].arXiv:1704.04861,2017.
[15] VIOLA P,JONES M J.Robust Real-Time Face Detection[J].International Journal of Computer Vision,2004,57(2):137-154.
[16] YAN J X,YIN J X.Multi-pose face detection based on face feature and AdaBoost algorithm[J].Journal of Computer Applications,2010,30(4):967-970.
[17] AHONEN T,HADID A,PIETIKAINEN M.Face Description with Local Binary Patterns:Application to Face Recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2006,28(12):2037-2041.
[18] YANG B,YAN J,LEI Z,et al.Aggregate channel features for multi-view face detection[C]//IEEE International Joint Confe-rence on Biometrics.2014:1-8.
[19] GHIASI G,FOWLKES C C.Occlusion Coherence:LocalizingOccluded Faces with a Hierarchical Deformable Part Model[C]//Computer Vision & Pattern Recognition.2014.
[20] LI Q Y,JI Q G,HONG S D.FastFace:Real-time robust face detection algorithm[J].Journal of Image and Graphics,2019,24(10):1761-1771.
[21] ZHANG K,ZHANG Z,LI Z,et al.Joint Face Detection andAlignment Using Multitask Cascaded Convolutional Networks[J].IEEE Signal Processing Letters,2016,23(10):1499-1503.
[22] YANG S,LUO P,LOY C C,et al.Faceness-Net:Face Detection through Deep Facial Part Responses[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,40(8):1845-1859.
[23] ZHOU X Y,WANG W,LI L Y.Overview of Target Detection Algorithm Based on Deep Learning[J].Electronic Measurement Technology,2017,40(11):89-93.
[24] MARCETIC D,HRKAC T,Slobodan Ribaric.Two-stage cascade model for unconstrained face detection[C]//2016 First International Workshop on Sensing,Processing and Learning for Intelligent Machines (SPLINE).IEEE,2016.
[25] ZHANG J L,WU X W,ZHU J K.Feature Agglomeration Networks for Single Stage Face Detection[J].arxiv:1712.00721,2017.
[26] LI E,WEN Z S,ZHANG Y L,et al.Face recognition systembased on OpenCV and mean hash algorithm[J].Industrial Design,2016(8):132-133.
[27] WANG A C,CHEN H T.Robust Perceptual Image Hashing Technique Based on Discrete Cosine Transform[J].China Safety Science Journal,2009,19(4):91.
[28] WANG C,KAO W H,HSIAO C K.Using Hamming Distance as Information for SNP-Sets Clustering and Testing in Disease Association Studies[J].Plos One,2015,10(8):e0135918.
[29] REN J J,WANG N.Study on the Loss Function in Artificial Neural Networks[J].Journal of Gansu Normal University,2018,23(2):65-67.
[30] RIBEIRO R,FEITOSA A R S,DE SOUZA R E,et al.Reconstruction of electrical impedance tomography images using genetic algorithms and non-blind search[C]//2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI 2014).IEEE,2014.
[31] ZHANG W X,LIANG Y.The Mathematical Foundation of Genetic Algorithm[J].Journal of Xi’an Jiaotong University,2000,34(10):6.
[32] GAO H K,MA S N.Research on Fast Face AuthenticationTechnology Based on ORL Database[J].Science & Technology Communication,2014,16(12):89-92.
[33] NGUYEN H V,BAI L.Cosine Similarity Metric Learning for Face Verification[C]//Computer Vision-ACCV 2010-10th Asian Conference on Computer Vision.Queenstown,New Zealand,2010.
[34] RAGHAVENDRA R,VENKATESH S,RAJA K B,et al.Mutual Information Based Multispectral Image Fusion for Improved Face Recognition[C]//2016 12th International Conference on Signal-Image Technology & Internet-Based Systems (SITIS).IEEE,2016.
[1] 郦睿翔, 毛莺池, 郝帅. 基于近似匹配的移动边缘计算缓存管理方法[J]. 计算机科学, 2021, 48(1): 96-102.
[2] 王瑞平, 贾真, 刘畅, 陈泽威, 李天瑞. 基于DeepFM的深度兴趣因子分解机网络[J]. 计算机科学, 2021, 48(1): 226-232.
[3] 于文家, 丁世飞. 基于自注意力机制的条件生成对抗网络[J]. 计算机科学, 2021, 48(1): 241-246.
[4] 仝鑫, 王斌君, 王润正, 潘孝勤. 面向自然语言处理的深度学习对抗样本综述[J]. 计算机科学, 2021, 48(1): 258-267.
[5] 丁钰, 魏浩, 潘志松, 刘鑫. 网络表示学习算法综述[J]. 计算机科学, 2020, 47(9): 52-59.
[6] 何鑫, 许娟, 金莹莹. 行为关联网络:完整的变化行为建模[J]. 计算机科学, 2020, 47(9): 123-128.
[7] 叶亚男, 迟静, 于志平, 战玉丽, 张彩明. 基于改进CycleGan模型和区域分割的表情动画合成[J]. 计算机科学, 2020, 47(9): 142-149.
[8] 暴雨轩, 芦天亮, 杜彦辉. 深度伪造视频检测技术综述[J]. 计算机科学, 2020, 47(9): 283-292.
[9] 袁野, 和晓歌, 朱定坤, 王富利, 谢浩然, 汪俊, 魏明强, 郭延文. 视觉图像显著性检测综述[J]. 计算机科学, 2020, 47(7): 84-91.
[10] 王文刀, 王润泽, 魏鑫磊, 漆云亮, 马义德. 基于堆叠式双向LSTM的心电图自动识别算法[J]. 计算机科学, 2020, 47(7): 118-124.
[11] 刘燕, 温静. 基于注意力机制的复杂场景文本检测[J]. 计算机科学, 2020, 47(7): 135-140.
[12] 张志扬, 张凤荔, 谭琪, 王瑞锦. 基于深度学习的信息级联预测方法综述[J]. 计算机科学, 2020, 47(7): 141-153.
[13] 蒋文斌, 符智, 彭晶, 祝简. 一种基于4Bit编码的深度学习梯度压缩算法[J]. 计算机科学, 2020, 47(7): 220-226.
[14] 陈晋音, 张敦杰, 林翔, 徐晓东, 朱子凌. 基于影响力最大化策略的抑制虚假消息传播的方法[J]. 计算机科学, 2020, 47(6A): 17-23.
[15] 程哲, 白茜, 张浩, 王世普, 梁宇. 使用深层卷积神经网络提高Hi-C 数据分辨率[J]. 计算机科学, 2020, 47(6A): 70-74.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 雷丽晖,王静. 可能性测度下的LTL模型检测并行化研究[J]. 计算机科学, 2018, 45(4): 71 -75 .
[2] 刘博艺,唐湘滟,程杰仁. 基于多生长时期模板匹配的玉米螟识别方法[J]. 计算机科学, 2018, 45(4): 106 -111 .
[3] 耿海军,施新刚,王之梁,尹霞,尹少平. 基于有向无环图的互联网域内节能路由算法[J]. 计算机科学, 2018, 45(4): 112 -116 .
[4] 崔琼,李建华,王宏,南明莉. 基于节点修复的网络化指挥信息系统弹性分析模型[J]. 计算机科学, 2018, 45(4): 117 -121 .
[5] 王振朝,侯欢欢,连蕊. 抑制CMT中乱序程度的路径优化方案[J]. 计算机科学, 2018, 45(4): 122 -125 .
[6] 韩奎奎,谢在鹏,吕鑫. 一种基于改进遗传算法的雾计算任务调度策略[J]. 计算机科学, 2018, 45(4): 137 -142 .
[7] 郑秀林,宋海燕,付伊鹏. MORUS-1280-128算法的区分分析[J]. 计算机科学, 2018, 45(4): 152 -156 .
[8] 吴姝,周安民,左政. PDiOS:iOS应用程序中私有API的调用检测[J]. 计算机科学, 2018, 45(4): 163 -168 .
[9] 朱淑芹,王文宏,李俊青. 针对基于感知器模型的混沌图像加密算法的选择明文攻击[J]. 计算机科学, 2018, 45(4): 178 -181 .
[10] 郭帅,刘亮,秦小麟. 用户偏好约束的空间关键词范围查询处理方法[J]. 计算机科学, 2018, 45(4): 182 -189 .