权值融合原始和虚拟样本的人脸识别方法

doi:10.3969/j.issn.1006-2475.2017.03.019

计算机与现代化

权值融合原始和虚拟样本的人脸识别方法

湖南师范大学物理与信息科学学院，湖南长沙410081

收稿日期:2016-08-25 出版日期:2017-03-29 发布日期:2017-03-30
作者简介:陈湘（1991-），女，湖南岳阳人，湖南师范大学物理与信息科学学院硕士研究生，研究方向：图像处理，模式识别；王玲（1962-），女，湖南长沙人，教授，硕士生导师，博士，研究方向：数字图像处理，现代通信与网络技术。

Face Recognition Method by Weight Fusion of Original and Mirror Samples

College of Physics and Information Science, Hunan Normal University, Changsha 410081, China

Received:2016-08-25 Online:2017-03-29 Published:2017-03-30

摘要/Abstract

摘要：

针对小样本问题（Small Sample Size Problem），提出一种基于局部二值模式（Local binary patterns,LBP）的权值融合原始和虚拟样本的人脸识别方法。首先，通过几何变换构造原始训练样

本的镜像虚拟样本；然后对所有训练样本进行分块，计算每一块的LBP特征直方图并串联起来作为一个样本的特征，再通过权值融合机制将原始样本和虚拟样本的特征合并成新的样本；最后，利用直方图

相交的方法进行判决分类。实验结果表明，当每类的训练样本数分别为1,2和3时，本算法在ORL和FERET人脸库上的识别率相较实验中的对比算法，分别提高了25.17%，15.8%和16.87%，取得了更高的识别

率。

关键词: 人脸识别, 小样本, 虚拟样本, 局部二值模式, 权值融合

Abstract:

Focused on the small sample size problem, a method for face recognition based on local binary pattern features of weight fusing the original and mirror image was

proposed. Firstly the proposed scheme generated the mirror images of the original training samples through the geometric transformation. Then each sample was divided into

several blocks, and the concatenated LBP histogram calculated over each block was used as the face feature of the sample. Next, the original and their mirror images face

feature was merged into new samples by weighted fusion scheme. Finally, the test sample was classified into the correct subject by using the histogram intersection method.

Experimental results showed that the algorithm has achieved a higher recognition rate, with the fact the recognition rate on FERET face database increased by 25.17%, 15.8% and

25.17%, respectively, compared with the contrast algorithm when the number of training sample per subject was 1,2 and 3 respectively. 

Key words: face recognition, small sample size, virtual sample, local binary pattern, weight fusion

中图分类号:

TP391.41

陈湘，王玲. 权值融合原始和虚拟样本的人脸识别方法[J]. 计算机与现代化, doi: 10.3969/j.issn.1006-2475.2017.03.019.

CHEN Xiang， WANG Ling. Face Recognition Method by Weight Fusion of Original and Mirror Samples[J]. Computer and Modernization, doi: 10.3969/j.issn.1006-2475.2017.03.019.

［1］Tan Xiaoyang, Chen Songcan, Zhou Zhihua, et al. Face recognition from a single image per person: A survey［J］. Pattern Recognition, 2006,39（9）:1725-1745. 
［2］Tang Bin, Luo Siwei, Huang Hua. High performance face recognition system by creating virtual sample［C］// Proceedings of International Conference on Neural Networks and

Signal Processing. 2003,2:972-975.
［3］Thian N P H, Marcel S, Bengio S. Improving face authentication using virtual samples［C］// Proceedings of the IEEE International Conference on Acoustics, Speech, and

Signal Processing. 2003,3:III233-236.
［4］Xu Yong, Zhu Xingjie, Li Zhengming, et al. Using the original and ‘symmetrical face training samples to perform representation based twostep face recognition ［J］.

Pattern Recognition, 2013,46（4）:1151-1158.
［5］Xu Yong, Fang Xiaozhao, Li Xuelong, et al. Data uncertainty in face recognition［J］. IEEE Transactions on Cybernentics, 2014,44（10）:1950-1961.
［6］Xu Yong, Li Xuelong, Yang Jian, et al. Integrate the original face image and its mirror image for face recognition［J］. Neurocomputing, 2014,131（7）:191-199.
［7］Ojala T, Pietikanten M, Hawood D. A comparative study of texture measure with classification based on feature distributions ［J］. Pattern Recognition, 1999,29（1）:51-59.


［8］Ojala T, Pietikainen M, Member S, et al. Multiresolution grayscale and rotation invariant texture classification with local binary patterns［J］. IEEE Transactions on

Pattern Analysis and Machine Intelligence, 2002,24（7）:971-987.
［9］Timo Ahonen, Abdenour Hadid, Matti Pietikainen. Face recognition with local binary patterns ［C］// Proceedings of the 8th European Conference on Computer Vision. 2004:469

-481.
［10］Zhang Jian, Xiao Xiao. Face recognition algorithm based on multilayer weighted LBP［C］// International Symposium on Computational Intelligence and Design. 2015:196-

199.
［11］兰佩,方超. 基于全局与局部特征融合的人脸识别方法［J］. 计算机与现代化, 2014（3）:109-113.
［12］Suruliandi A, Meena K, Rose Reena R. Local binary pattern and its derivatives for face recognition［J］. IET Computer Vision, 2012,6（5）:480-488.
［13］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．
［14］李昕昕,陈丹,许凤娇. 面向光照可变的人脸识别方法［J］. 计算机应用, 2013,33（2）:507-510,514.
［15］Xu Yong, Jin Zhong. Downsampling face images and lowresolution face recognition［C］// The 3rd International Conference on Innovative Computing Information and

Control. 2008:392-395.

[1]	夏千涵, 何胜煌, 吴元清, 赵乐乐. 基于可学习记忆特征金字塔网络的小样本目标检测[J]. 计算机与现代化, 2023, 0(12): 7-13.
[2]	王鑫, 肖韬睿. 基于生成对抗网络的人脸识别对抗攻击[J]. 计算机与现代化, 2023, 0(10): 115-120.
[3]	于韬, 张英, 拥措, . 基于小样本学习的藏文命名实体识别[J]. 计算机与现代化, 2023, 0(05): 13-19.
[4]	盛江岸, 陈淑荣. 融合双重注意力机制的戴口罩人脸识别方法[J]. 计算机与现代化, 2023, 0(02): 72-77.
[5]	王伟, 余厚云, . 基于改进深度度量学习算法的表面缺陷检测[J]. 计算机与现代化, 2021, 0(06): 61-68.
[6]	徐龙. 基于自注意力和胶囊网络的短文本情感分析[J]. 计算机与现代化, 2020, 0(07): 61-64.
[7]	陈凯,祖莉,欧屹. 基于YOLOv3与ResNet50的摄影机器人人脸识别跟踪系统[J]. 计算机与现代化, 2020, 0(04): 30-.
[8]	樊笛,巨志勇. 基于原型网络的小样本图像识别方法[J]. 计算机与现代化, 2020, 0(03): 103-.
[9]	李硕1，卞青山2，刘传文2，刘鸣涛1，张林涛1 . 基于DCNN的证件照人脸验证及应用研究[J]. 计算机与现代化, 2020, 0(02): 104-.
[10]	崔娟娟1，张蕾1,侯谢炼2，陈才扣2，张海燕1. 基于鉴别性低秩表示的2阶段人脸识别算法[J]. 计算机与现代化, 2019, 0(12): 55-.
[11]	李垒，任越美. 基于多任务联合判别稀疏表示的人脸识别[J]. 计算机与现代化, 2019, 0(10): 72-.
[12]	黄武飞1，曾吉昌1，黄锦权1，李思寒1，陈育均1，陈钊1，郑志硕1,2. 基于实验室安保服务型机器人的人脸识别[J]. 计算机与现代化, 2019, 0(07): 61-.
[13]	郑健,王志明,张宁. 一种基于深度学习的改进人脸识别算法[J]. 计算机与现代化, 2018, 0(12): 90-.
[14]	潘峥嵘，王震. 一种结合Gabor小波与深度学习的人脸识别方法[J]. 计算机与现代化, 2018, 0(11): 46-.
[15]	郑秋梅，谢换丽，王风华，苏政，刘真. 一种基于局部三值模式的深度学习人脸识别算法[J]. 计算机与现代化, 2018, 0(02): 112-.

权值融合原始和虚拟样本的人脸识别方法

Face Recognition Method by Weight Fusion of Original and Mirror Samples

可视化

被引次数

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价