|本期目录/Table of Contents|

[1]陈希彤,卢 涛*.基于全局深度分离卷积残差网络的高效人脸识别算法[J].武汉工程大学学报,2019,(03):276-282.[doi:10. 3969/j. issn. 1674-2869. 2019. 03. 014]
 CHEN Xitong,LU Tao *.Efficient Face Recognition Algorithm Using Global Deep Separable Convolutional and Residual Network[J].Journal of Wuhan Institute of Technology,2019,(03):276-282.[doi:10. 3969/j. issn. 1674-2869. 2019. 03. 014]
点击复制

基于全局深度分离卷积残差网络的高效人脸识别算法(/HTML)
分享到:

《武汉工程大学学报》[ISSN:1674-2869/CN:42-1779/TQ]

卷:
期数:
2019年03期
页码:
276-282
栏目:
机电与信息工程
出版日期:
2019-06-20

文章信息/Info

Title:
Efficient Face Recognition Algorithm Using Global Deep Separable Convolutional and Residual Network
文章编号:
20190314
作者:
陈希彤卢 涛*
武汉工程大学计算机科学与工程学院,湖北 武汉 430205
Author(s):
CHEN Xitong LU Tao *
School of Computer Science and Engineering,WuhanInstitute of Technology, Wuhan 430205, China
关键词:
人脸识别可分离卷积残差学习卷积神经网络
Keywords:
face recognition separable convolution residual learning convolutional neural network
分类号:
TP391.4
DOI:
10. 3969/j. issn. 1674-2869. 2019. 03. 014
文献标志码:
A
摘要:
深度学习模型的复杂性影响了人脸识别的实时性能,限制了人脸识别算法在实际场景中的应用。针对这一问题,提出了一种基于全局深度分离卷积的残差学习神经网络,首先利用小卷积核提取人脸图像局部细节信息,采用深度残差学习网络作为骨干网络提取不同层次特征,然后根据人脸特征分布的空间重要性使用全局深度可分离卷积调整学习权重,加速精炼深层抽象特征,通过这一机制获取判别能力更强的特征向量进行人脸识别。在CASIA-Webface与Extend Yale-B人脸数据集中的识别率分别达到了82.1%与99.8%。
Abstract:
Because the model complexity of deep learning affects the real-time performance of face recognition, this limits the wide application of face recognition algorithm in real-world scenario. To address this problem, we proposed a residual neural network by global deep separable convolution. Firstly, we extracted local details of face images by using a small convolution kernel, then the deep residual learning network was used as the backbone network for extracting different levels of features. According to the spatial importance of face feature distribution, we used the global deep separable convolution to adjust learn weights, accelerate and refine deep features. The mechanism obtains more discriminative feature vector for face recognition. The recognition rates of CASIA-Web face and Extended Yale-B face datasets reach 82.1% and 99.8%, respectively.

参考文献/References:

[1] KRIZHEVSKY  A,  SUTSKEVER I, HINTON G E. Imagenet classification with deep convolutional neural networks[C]//Advances in Neural Information Processing Systems. Lake Tahoe:ACM, 2012: 1097- 1105. [2] 卢涛, 章瑾, 陈白帆,等.多尺度自适应配准的视频超分辨率算法[J]. 武汉工程大学学报, 2016, 38(2):178-184. [3] 夏平平, 吕太之. 动态人脸识别系统的设计与实现[J]. 武汉工程大学学报, 2011, 33(10):107-110. [4] 赵建民,朱信忠,江小辉.基于改进型LBP特征的人脸识别方法研究[J]. 计算机科学,2009,36(8):276-280. [5] 王亚星, 齐林, 郭新, 等. 基于稀疏 PCA 的多阶次分数阶傅里叶变换域特征人脸识别[J]. 计算机应用研究, 2016, 33(4): 1253-1257. [6] 杨威,卢涛,汪浩. 基于半耦合稀疏表达的极低分辨率人脸识别[J]. 计算机工程与应用, 2017, 53(22): 169-175. [7] 赵娜, 赵彤洲, 邹冲,等. 稀疏表示中字典学习的影响因子研究[J]. 武汉工程大学学报, 2017, 39(3):267-272. [8] HINTON G E, SALAKUTDINOV R R. Reducing the dimensionality of data with neural networks[J]. Science, 2006, 313(5786): 504-507. [9] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[C]//International Conference on Learning Representations. San Diego:ICLR,2015. [10] HE  K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision And Pattern Recognition. Las Vegas: IEEE, 2016: 770-778. [11] HUANG G, LIU Z, VAN DER MAATEN L, et al. Densely connected convolutional networks[C]//Proceedings of the IEEE Conference on Computer Vision And Pattern Recognition. Hawaii: IEEE, 2017: 4700-4708. [12] CHEN S, LIU Y, GAO X, et al. Mobilefacenets: efficient cnns for accurate real-time face verification on mobiledevices[C]//Chinese Conference on Biometric Recognition. Cham:Springer, 2018: 428-438. [13] SZEGEDY C, VANHOUCKEV, IOFFE S, et al. Rethinking the inception architecture for computer vision[C]//Proceedings of the IEEE Conference on Computer Vision And Pattern Recognition. Las Vegas: IEEE, 2016: 2818-2826. [14] DUVENAUD D, RIPPEL O, Adams R, et al. Avoiding pathologies in very deep networks[C]//Artificial Intelligence and Statistics. Reykjavik: JMLR,2014: 202-210. [15] WU X, HE R, SUN Z A, et al. A light CNN for deep face representation with noisy labels[J]. IEEE Transactions on Information Forensics and Security, 2018, 13(11):2884-2896. [16] DE BOER P T, KROESE D P, MANNOR S, et al. A tutorial on the cross-entropy method[J]. Annals of operations research, 2005, 134(1): 19-67.

相似文献/References:

[1]夏平平,吕太之.动态人脸识别系统的设计与实现[J].武汉工程大学学报,2011,(10):107.
 XIA Ping ping,LV Tai zhi.Design and implementation of dynamic faces recognition system[J].Journal of Wuhan Institute of Technology,2011,(03):107.
[2]韦 仙,康睿丹.基于降维压缩法的图像重构[J].武汉工程大学学报,2015,37(12):69.[doi:10. 3969/j. issn. 1674-2869. 2015. 12. 015]
 -.Image reconstruction based on dimension reduction and compression technology[J].Journal of Wuhan Institute of Technology,2015,37(03):69.[doi:10. 3969/j. issn. 1674-2869. 2015. 12. 015]
[3]郭 婷,张天序*,郭诗嘉.一种红外图像和宽光谱融合的人脸识别算法[J].武汉工程大学学报,2022,44(03):320.[doi:10.19843/j.cnki.CN42-1779/TQ.202111015]
 GUO Ting,ZHANG Tianxu*,GUO Shijia.Face Recognition Algorithm Based on Fusion of Infrared Image and Wide Spectrum Features[J].Journal of Wuhan Institute of Technology,2022,44(03):320.[doi:10.19843/j.cnki.CN42-1779/TQ.202111015]

备注/Memo

备注/Memo:
收稿日期:2019-03-02基金项目:国家自然科学基金(61502354,61671332,41501505);湖北省自然科学基金(2015CFB451,2014CFA130,2012FFA099,2012FFA134,2013CF125)作者简介:陈希彤, 硕士研究生。 E-mail:375122362@qq.com*通讯作者:卢 涛, 博士, 副教授。 E-mail:lut@wit.edu.cn引文格式:陈希彤,卢涛. 基于全局深度分离卷积残差网络的高效人脸识别算法[J]. 武汉工程大学学报,2019,41(3):276-282.
更新日期/Last Update: 2019-06-19