一种用于多源人脸图像检索的损失函数改进策略

任国印; 吕晓琪; 李宇豪

doi:10.37188/CJLCD.2021-0003

您当前的位置：

首页 >

文章列表页 >

一种用于多源人脸图像检索的损失函数改进策略

目标识别与成像 | 更新时间：2021-11-08

- 一种用于多源人脸图像检索的损失函数改进策略
- Improved loss function strategy for multi source face image retrieval
- 液晶与显示 2021年36卷第11期页码：1583-1596
- 作者机构：
  
  1.内蒙古科技大学机械工程学院, 内蒙古包头 014010
  2.内蒙古科技大学信息工程学院, 内蒙古包头 014010
  3.内蒙古工业大学, 内蒙古呼和浩特 010051
- 作者简介：
  
  [ "任国印(1985-), 男, 内蒙古包头人, 博士研究生, 讲师, 2013年于内蒙古科技大学获得硕士学位, 主要从事模式识别和智能图像处理方面的研究, E-mail: 1712152231@qq.com" ]
  [ "吕晓琪(1963-), 男, 内蒙古包头人, 博士, 教授, 2003年于北京科技大学获得博士学位, 主要从事模式识别和智能图像处理的研究。E-mail: 635302395@qq.com" ]
- 基金信息：
  
  国家自然科学基金(61179019);国家自然科学基金(81571753);CERNET Innovation Project(NGII20170705);包头市青年创新人才项目(0701011904)
- DOI：10.37188/CJLCD.2021-0003
  中图分类号： TP391.4
- 收稿日期：2021-01-05，
  
  修回日期：2021-04-09，
  
  纸质出版日期：2021-11
- 稿件说明：
移动端阅览
任国印, 吕晓琪, 李宇豪. 一种用于多源人脸图像检索的损失函数改进策略[J]. 液晶与显示, 2021,36(11):1583-1596.

Guo-yin REN, Xiao-qi LYU, Yu-hao LI. Improved loss function strategy for multi source face image retrieval[J]. Chinese journal of liquid crystals and displays, 2021, 36(11): 1583-1596.
任国印, 吕晓琪, 李宇豪. 一种用于多源人脸图像检索的损失函数改进策略[J]. 液晶与显示, 2021,36(11):1583-1596. DOI： 10.37188/CJLCD.2021-0003.

Guo-yin REN, Xiao-qi LYU, Yu-hao LI. Improved loss function strategy for multi source face image retrieval[J]. Chinese journal of liquid crystals and displays, 2021, 36(11): 1583-1596. DOI： 10.37188/CJLCD.2021-0003.

摘要

目前，公安机关刑侦部门大量的离线视频存储在监控网络的服务器中，为了在这些海量视频帧中提取嫌疑人目标人脸，设计了人脸检索系统。通过改变CNN网络的RELU结构，训练新的四分支孪生网络来获取深度特征，构造了一个新的四分支孪生网络。结合网上发布的逃犯人脸图片发起通缉，借助深度特征对比展开基于内容的图像检索（CBIR）。新的四分支孪生网络比熟悉的网络，如Alexnet、Googlenet、VGGNet和ResNet等收敛得更快，系统鲁棒性好。网络的共享权重设计使得检索具有较高的模型训练精度和检索精度。图像深度特征可以在摄像机之间快速在线共享。实验结果表明，该方法的平均检索精度（ARP）为98.74%，模型训练精度为99.51%，帧率为28 FPS。

Abstract

At present

a large number of offline videos are stored in the servers of the surveillance network in criminal investigation departments of public security organs

and the face retrieval system is designed. The new Quadruplet Network converges faster than the familiar networks such as Alexnet

Googlenet

VGGNet and ResNet. Because of the shared weight design of the network

the retrieval has a high precision

Average Retrieval Precision(ARP) and model training accuracy

and the system has good robustness. The image depth features can be shared quickly online between the cameras. The proposed method is effective

with ARP of 98.74% and a model training accuracy of 99.51%

and a frame rate of 28 FPS.

关键词

Keywords

references

WU H, YOKOYAMA T, PRAMADIHANTO D, et al . Face and facial feature extraction from color image[C]// Proceedings of the Second International Conference on Automatic Face and Gesture Recognition . Killington, VT, USA: IEEE, 1996: 345-350.

SHI Y X, REN X X, YANG S Z, et al . A generalized Kernel Fisher Discriminant framework used for feature extraction and face recognition[C]// Proceedings of the 2016 12 th International Conference on Natural Computation , Fuzzy Systems and Knowledge Discovery ( ICNC - FSKD ). Changsha, China: IEEE, 2016: 1487-1491.

Z LU , X D JIANG , A KOT . Deep coupled ResNet for low-resolution face recognition . IEEE Signal Processing Letters , 2018 . 25 ( 4 ): 526 - 530 . DOI: 10.1109/LSP.2018.2810121 http://doi.org/10.1109/LSP.2018.2810121 .

C SCHMID , R MOHR . Local grayvalue invariants for image retrieval . IEEE Transactions on Pattern Analysis and Machine Intelligence , 1997 . 19 ( 5 ): 530 - 535 . DOI: 10.1109/34.589215 http://doi.org/10.1109/34.589215 .

A RAZA , H DAWOOD , H DAWOOD , 等 . Correlated primary visual Texton histogram features for content base image retrieval . IEEE Access , 2018 . 6 46595 - 46616 . DOI: 10.1109/ACCESS.2018.2866091 http://doi.org/10.1109/ACCESS.2018.2866091 .

J XIANG , N ZHANG , R R PAN , 等 . Fabric image retrieval system using hierarchical search based on deep convolutional neural network . IEEE Access , 2019 . 7 35405 - 35417 . DOI: 10.1109/ACCESS.2019.2898906 http://doi.org/10.1109/ACCESS.2019.2898906 .

T T D PHAM , S KIM , Y C LU , 等 . Facial action units-based image retrieval for facial expression recognition . IEEE Access , 2019 . 7 5200 - 5207 . DOI: 10.1109/ACCESS.2018.2889852 http://doi.org/10.1109/ACCESS.2018.2889852 .

Z DONG , C C JING , M T PEI , 等 . Deep CNN based binary hash video representations for face retrieval . Pattern Recognition , 2018 . 81 357 - 369 . DOI: 10.1016/j.patcog.2018.04.014 http://doi.org/10.1016/j.patcog.2018.04.014 .

S NAGPAL , M SINGH , R SINGH , 等 . Regularized deep learning for face recognition with weight variations . IEEE Access , 2015 . 3 3010 - 3018 . DOI: 10.1109/ACCESS.2015.2510865 http://doi.org/10.1109/ACCESS.2015.2510865 .

M A ABUZNEID , A MAHMOOD . Enhanced human face recognition using LBPH descriptor, Multi-KNN, and back-propagation neural network . IEEE Access , 2018 . 6 20641 - 20651 . DOI: 10.1109/ACCESS.2018.2825310 http://doi.org/10.1109/ACCESS.2018.2825310 .

肖康 , 田立君 , 王中阳 . 基于深度学习的低信噪比下的快速超分辨荧光显微成像 . 中国激光 , 2020 . 47 ( 10 ): 222 - 232 . https://www.cnki.com.cn/Article/CJFDTOTAL-JJZZ202010027.htm https://www.cnki.com.cn/Article/CJFDTOTAL-JJZZ202010027.htm .

K XIAO , L J TIAN , Z Y WANG . Fast super-resolution fluorescence microscopy imaging with low signal-to-noiseratio based on deep learning . Chinese Journal of Lasers , 2020 . 47 ( 10 ): 222 - 232 . https://www.cnki.com.cn/Article/CJFDTOTAL-JJZZ202010027.htm https://www.cnki.com.cn/Article/CJFDTOTAL-JJZZ202010027.htm .

胡麟苗 , 张湧 , 楼晨风 . 基于内容特征提取的短波红外-可见光人脸识别 . 光学精密工程 , 2021 . 29 ( 1 ): 160 - 171 . https://www.cnki.com.cn/Article/CJFDTOTAL-GXJM202101017.htm https://www.cnki.com.cn/Article/CJFDTOTAL-GXJM202101017.htm .

L M HU , Y ZHANG , C F LOU . Shortwave infrared visible-light face recognition based on content feature extraction . Optics and Precision Engineering , 2021 . 29 ( 1 ): 160 - 171 . https://www.cnki.com.cn/Article/CJFDTOTAL-GXJM202101017.htm https://www.cnki.com.cn/Article/CJFDTOTAL-GXJM202101017.htm .

明悦 , 王绍颖 , 范春晓 , 等 . 对齐特征表示的跨模态人脸识别 . 光学精密工程 , 2020 . 28 ( 10 ): 2311 - 2322 . DOI: 10.37188/OPE.20202810.2311 http://doi.org/10.37188/OPE.20202810.2311 .

Y MING , S Y WANG , C X FAN , 等 . Exploring aligned latent representations for cross-domain face recognition . Optics and Precision Engineering , 2020 . 28 ( 10 ): 2311 - 2322 . DOI: 10.37188/OPE.20202810.2311 http://doi.org/10.37188/OPE.20202810.2311 .

秦智康 , 张衡阳 . 基于变换域中的自适应纹理图像检索 . 液晶与显示 , 2020 . 35 ( 3 ): 281 - 286 . DOI: 10.3788/YJYXS20203503.0281 http://doi.org/10.3788/YJYXS20203503.0281 .

Z K QIN , H Y ZHANG . Adaptive texture image retrieval based on transform domain . Chinese Journal of Liquid Crystals and Displays , 2020 . 35 ( 3 ): 281 - 286 . DOI: 10.3788/YJYXS20203503.0281 http://doi.org/10.3788/YJYXS20203503.0281 .

SUN Y, CHEN Y H, WANG X G, et al . Deep learning face representation by joint identification-verification[C]// Proceedings of the 27 th International Conference on Neural Information Processing Systems . Montreal, Canada: MIT Press, 2014: 1988-1996.

M M ZHANG , W LI , Q DU . Diverse region-based CNN for hyperspectral image classification . IEEE Transactions on Image Processing , 2018 . 27 ( 6 ): 2623 - 2634 . DOI: 10.1109/TIP.2018.2809606 http://doi.org/10.1109/TIP.2018.2809606 .

Y D ZHAO , K R HAO , H B HE , 等 . A visual long-short-term memory based integrated CNN model for fabric defect image classification . Neurocomputing , 2020 . 380 259 - 270 . DOI: 10.1016/j.neucom.2019.10.067 http://doi.org/10.1016/j.neucom.2019.10.067 .

SARDOGAN M, TUNCER A, OZEN Y. Plant leaf disease detection and classification based on CNN with LVQ algorithm[C]//2018 3 rd International Conference on Computer Science and Engineering ( UBMK ). Sarajevo, Bosnia-Herzegovina: IEEE, 2018: 382-385.

GE W F, HUANG W L, DONG D K, et al . Deep metric learning with hierarchical triplet loss[C]// Proceedings of the European Conference on Computer Vision ( ECCV ). Munich, Germany: Springer, 2018: 272-288.

CHEN W H, CHEN X T, ZHANG J G, et al . Beyond triplet loss: A deep quadruplet network for person re-identification[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition ( CVPR ). Honolulu, HI, USA: IEEE, 2017: 1320-1329.

LV X B, ZHAO C R, CHEN W. A novel hard mining center-triplet loss for person re-identification[C]// Chinese Conference on Pattern Recognition and Computer Vision ( PRCV ). Xi'an, China: Springer, 2019: 199-210.

H P YAO , D Y FU , P Y ZHANG , 等 . MSML: A novel multilevel semi-supervised machine learning framework for intrusion detection system . IEEE Internet of Things Journal , 2019 . 6 ( 2 ): 1949 - 1959 . DOI: 10.1109/JIOT.2018.2873125 http://doi.org/10.1109/JIOT.2018.2873125 .

SUN Y, LIANG D, WANG X G, et al . DeepID3: Face recognition with very deep neural networks[J]. arXiv preprint arXiv : 1502.00873, 2015.

LIN T Y, GOYAL P, GIRSHICK R, et al . Focal loss for dense object detection[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision ( ICCV ). Venice, Italy: IEEE, 2017: 2999-3007.

WANG C, LAN X P, ZHANG X. How to train triplet networks with 100 K identities?[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision Workshops ( ICCVW ). Venice, Italy: IEEE, 2017: 1907-1915.

H J ZHANG , Y Z JI , W HUANG , 等 . Sitcom-star-based clothing retrieval for video advertising: a deep learning framework . Neural Computing and Applications , 2019 . 31 7361 - 7380 . DOI: 10.1007/s00521-018-3579-x http://doi.org/10.1007/s00521-018-3579-x .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

暂无数据