高级搜索

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于球面Haar小波和卷积神经网络的飞行员虹膜识别

贾博 冯孝鑫 李军 俞碧婷 赵倩 吴奇

贾博, 冯孝鑫, 李军, 俞碧婷, 赵倩, 吴奇. 基于球面Haar小波和卷积神经网络的飞行员虹膜识别[J]. 电子与信息学报, 2021, 43(4): 939-947. doi: 10.11999/JEIT190928
引用本文: 贾博, 冯孝鑫, 李军, 俞碧婷, 赵倩, 吴奇. 基于球面Haar小波和卷积神经网络的飞行员虹膜识别[J]. 电子与信息学报, 2021, 43(4): 939-947. doi: 10.11999/JEIT190928
Bo JIA, Xiaoxin FENG, Jun LI, Biting YU, Qian ZHAO, Qi WU. Pilot Iris Recognition Based on Spherical Haar Wavelet and Convolutional Neural Network[J]. Journal of Electronics & Information Technology, 2021, 43(4): 939-947. doi: 10.11999/JEIT190928
Citation: Bo JIA, Xiaoxin FENG, Jun LI, Biting YU, Qian ZHAO, Qi WU. Pilot Iris Recognition Based on Spherical Haar Wavelet and Convolutional Neural Network[J]. Journal of Electronics & Information Technology, 2021, 43(4): 939-947. doi: 10.11999/JEIT190928

基于球面Haar小波和卷积神经网络的飞行员虹膜识别

doi: 10.11999/JEIT190928
基金项目: 国家自然科学基金(U1933125)
详细信息
    作者简介:

    贾博:男,1987年生,工程师,研究方向为航空大数据、民航安全

    冯孝鑫:男,1997年生,硕士生,研究方向为信号处理、计算机视觉

    李军:男,1968年生,一级飞行员,研究方向为民航运行、飞行理论

    俞碧婷:女,1990年生,博士,研究方向为机器视觉、深度学习

    赵倩:女,1994年生,飞行理论教员,研究方向为民航飞行教学

    吴奇:男,1978年生,副教授,研究方向为视脑交互

    通讯作者:

    吴奇 wuqi7812@sjtu.edu.cn

  • 中图分类号: TP181

Pilot Iris Recognition Based on Spherical Haar Wavelet and Convolutional Neural Network

Funds: The National Natural Science Foundation of China (U1933125)
  • 摘要: 虹膜识别面临两个重要的问题:一是如何精细分解与重构虹膜球面图像;二是如何识别虹膜图特征。虹膜表面几何位置信息是一种重要的信号,传统的虹膜识别通常使用虹膜图像的平面特征,然而人的眼睛是一种球体,从平面图像难以提取到虹膜球体的几何特征。针对平面特征容易出现虹膜纹理的扭曲和失真等问题,该文建议一种正交对称的球面Haar小波(OSSHW)基,对球面虹膜信号进行多尺度分解与重构,获得更精细的虹膜曲面几何特征,同时对比球谐函数和半正交或正交球面Haar小波基的虹膜球面信号特征提取能力。在此基础上,该文提出一种基于卷积神经网络(CNN)和正交对称的球面Haar小波的虹膜识别方法,它能够有效捕获虹膜球体曲面的局部精细特征,比半正交或正交球面Haar小波基具有更强的虹膜识别能力。
  • 图  1  仿真实验流程图

    图  2  球面谐波函数重构图像

    图  3  球面三角形划分方案

    图  4  虹膜信号重构误差

    图  5  检测得到虹膜内外边缘

    图  6  检测得到睫毛和上眼睑

    图  7  分离出的虹膜图像

    图  8  卷积神经网络结构图

    表  1  使用5种球面Haar小波基进行虹膜信号重构$ {l}_{2} $误差(保留1.56%的小波系数,level=5)

    小波基虹膜信号
    Bio HaarNielsonBonneauPseudo HaarOSSHW
    10.26710.24320.24480.25350.2282
    20.19170.17440.17170.18130.1690
    30.29670.27200.27160.28600.2676
    40.25190.23450.23400.24460.2310
    50.26360.24570.24830.25710.2432
    下载: 导出CSV

    表  2  使用不同球面信号分析方法的识别准确率

    网络结构准确率(CCR)(%)耗时(s)
    Dhage等人[14]97.234.88
    Bharath等人[15]95.930.10
    DeepIris[16]95.50/
    Chen等人[17]98.00/
    IrisConvShallower98.10/
    IrisConvDeeper98.79/
    球谐函数 + CNN91.074.36
    Bio Haar + CNN95.533.98
    Nielson + CNN96.423.34
    Bonneau + CNN97.323.27
    Pseudo Haar + CNN97.322.95
    OSSHW + CNN98.213.92
    下载: 导出CSV
  • MA Li, WANG Yunhong, and TAN Tieniu. Iris recognition based on multichannel Gabor filtering[C]. The 5th Asian Conference on Computer Vision, Melbourne, Australia, 2002: 279–283.
    YAO Peng, LI Jun, YE Xueyi, et al. Iris recognition algorithm using modified log-Gabor filters[C]. The 18th International Conference on Pattern Recognition (ICPR’06), Hong Kong, China, 2006: 461–464.
    NABTI M and BOURIDANE A. An effective and fast iris recognition system based on a combined multiscale feature extraction technique[J]. Pattern Recognition, 2008, 41(3): 868–879. doi: 10.1016/j.patcog.2007.06.030
    PROENÇA H and SANTOS G. Fusing color and shape descriptors in the recognition of degraded iris images acquired at visible wavelengths[J]. Computer Vision and Image Understanding, 2012, 116(2): 167–178. doi: 10.1016/j.cviu.2011.10.008
    SANTOS G and HOYLE E. A fusion approach to unconstrained iris recognition[J]. Pattern Recognition Letters, 2012, 33(8): 984–990. doi: 10.1016/j.patrec.2011.08.017
    TAN Tieniu, ZHANG Xiaobo, SUN Zhenan, et al. Noisy iris image matching by using multiple cues[J]. Pattern Recognition Letters, 2012, 33(8): 970–977. doi: 10.1016/j.patrec.2011.08.009
    HUO Guang, LIU Yuanning, ZHU Xiaodong, et al. Secondary iris recognition method based on local energy-orientation feature[J]. Journal of Electronic Imaging, 2015, 24(1): 013033. doi: 10.1117/1.JEI.24.1.013033
    KUMAR A, POTNIS A, and SINGH A. Iris recognition and feature extraction in iris recognition system by employing 2D DCT[J]. IRJET International Research Journal of Engineering and Technology, 2016, 3(12): 503–510.
    刘元宁, 刘帅, 朱晓冬, 等. 基于特征加权融合的虹膜识别算法[J]. 吉林大学学报: 工学版, 2019, 49(1): 221–229.

    LIU Yuanning, LIU Shuai, ZHU Xiaodong, et al. Iris recognition algorithm based on feature weighted fusion[J]. Journal of Jilin University:Engineering and Technology Edition, 2019, 49(1): 221–229.
    OUYANG Wanli, ZHAO Tianle, CHAM W K, et al. Fast full-search-equivalent pattern matching using asymmetric haar wavelet packets[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2016, 28(4): 819–833. doi: 10.1109/tcsvt.2016.2629621
    WU E Q, ZHOU Guirong, ZHU Limin, et al. Rotated sphere haar wavelet and deep contractive auto-encoder network with fuzzy Gaussian SVM for pilot’s pupil center detection[J]. IEEE Transactions on Cybernetics, 2019, 51(1): 332–345. doi: 10.1109/TCYB.2018.2886012
    XU Guangzhu, ZHANG Zaifeng, and MA Yide. A novel method for iris feature extraction based on intersecting cortical model network[J]. Journal of Applied Mathematics and Computing, 2018, 26(1/2): 341–352.
    ALOYSIUS N and GEETHA M. A review on deep convolutional neural networks[C]. 2017 International Conference on Communication and Signal Processing (ICCSP), Chennai, India, 2017: 588–592.
    DHAGE S S, HEGDE S S, MANIKANTAN K, et al. DWT-based feature extraction and radon transform based contrast enhancement for improved iris recognition[J]. Procedia Computer Science, 2015, 45: 256–265. doi: 10.1016/j.procs.2015.03.135
    BHARATH B V, VILAS A S, MANIKANTAN K, et al. Iris recognition using radon transform thresholding based feature extraction with Gradient-based Isolation as a pre-processing technique[C]. The 9th International Conference on Industrial and Information Systems (ICIIS), Gwalior, India, 2014: 1–8.
    MINAEE S, ABDOLRASHIDI A. Deepiris: Iris recognition using a deep learning approach[J]. arXiv preprint arXiv: 1907.09380, 2019.
    CHEN Ying, WANG Wenyuan, ZENG Zhuang, et al. Adapted deep convnets technology for robust iris recognition[J]. Journal of Electronic Imaging, 2019, 28(3): 033008.
  • 加载中
图(8) / 表(2)
计量
  • 文章访问数:  1581
  • HTML全文浏览量:  920
  • PDF下载量:  87
  • 被引次数: 0
出版历程
  • 收稿日期:  2019-11-20
  • 修回日期:  2021-01-15
  • 网络出版日期:  2021-01-22
  • 刊出日期:  2021-04-20

目录

    /

    返回文章
    返回