您的位置:山东大学 -> 科技期刊社 -> 《山东大学学报(工学版)》

山东大学学报 (工学版) ›› 2019, Vol. 49 ›› Issue (2): 54-60.doi: 10.6040/j.issn.1672-3961.0.2017.420

• 机器学习与数据挖掘 • 上一篇    下一篇

基于多重多元回归的人脸年龄估计

向润1(),陈素芬2,曾雪强3,*()   

  1. 1. 南昌大学信息工程学院,江西 南昌 330031
    2. 南昌工程学院信息工程学院,江西 南昌 330099
    3. 江西师范大学计算机信息工程学院,江西 南昌 330022
  • 收稿日期:2017-08-24 出版日期:2019-04-20 发布日期:2019-04-19
  • 通讯作者: 曾雪强 E-mail:xiangr214@foxmail.com;xqzeng@jxnu.edu.cn
  • 作者简介:向润(1993—),男,湖北黄冈人,硕士研究生,主要研究方向为机器学习.E-mail:xiangr214@foxmail.com
  • 基金资助:
    国家自然科学基金资助项目(61463033);国家自然科学基金资助项目(61866017);江西省杰出青年人才资助计划资助项目(20171BCB23013);江西省自然科学基金资助项目(20151BAB207028)

Facial age estimation based on multivariate multiple regression

Run XIANG1(),Sufen CHEN2,Xueqiang ZENG3,*()   

  1. 1. Information Engineering School, Nanchang University, Nanchang 330031, Jiangxi, China
    2. School of Information Engineering, Nanchang Institute of Technology, Nanchang 330099, Jiangxi, China
    3. School of Computer & Information Engineering, Jiangxi Normal University, Nanchang 330022, Jiangxi, China
  • Received:2017-08-24 Online:2019-04-20 Published:2019-04-19
  • Contact: Xueqiang ZENG E-mail:xiangr214@foxmail.com;xqzeng@jxnu.edu.cn
  • Supported by:
    国家自然科学基金资助项目(61463033);国家自然科学基金资助项目(61866017);江西省杰出青年人才资助计划资助项目(20171BCB23013);江西省自然科学基金资助项目(20151BAB207028)

摘要:

基于标记分布学习的人脸年龄估计方法利用相近年龄的人脸变化较为缓慢的特点,采用年龄标记分布向量表示附近年龄描述目标年龄的程度,将学习任务从单值的目标年龄预测转变为年龄标记分布向量的估计,较为有效的解决了人脸年龄估计任务中训练数据不足的问题。但是,现有的标记分布学习方法存在不能构建统一的标记分布预测模型(基于最大熵模型的方法)或容易过拟合的问题(基于神经网络的方法)。为了解决这些问题,将基于标记分布学习的年龄估计转换为同时对多因变量进行预测的多重多元回归分析问题,并采用多因变量偏最小二乘回归方法进行求解。多因变量偏最小二乘回归模型对数据分布没有前提假定,在自变量存在较大的相关性的情况下仍可建立有效的多因变量预测模型。在FG-NET人脸数据库上的大量对比试验结果表明,本研究提出的基于多重多元回归的人脸年龄估计方法在大幅度提高模型训练效率的同时,具有更高的年龄估计准确度。

关键词: 人脸年龄估计, 多重多元回归, 偏最小二乘回归, 标记分布学习, 最小二乘回归

Abstract:

Label distribution learning based facial age estimation model was an effective method to solve the problem of insufficient training data caused by the difficulty of facial image collection, where its motivation was that facial aging information on adjacent ages can be introduced to enhance the age estimation model due to human faces changing slowly. Given a certain age to learn, label distribution learning converted the learning target from a continuous value to an age label distribution vector, which was generated according to the description degree of the neighboring ages. However, the existed methods had the drawbacks of separated age prediction model (maximum entropy based methods) or tending to be overfitting (neural network based methods). So a method of facial age estimation based on multivariate multiple regression was proposed, the label distribution learning based age estimation problem was transformed into a multivariate multiple regression analysis task and then solved by the multivariate partial least squares regression. Multivariate partial least squares regression had no assumption about the data distribution and built an integrated effective model for all ages even when there is a strong correlation among independent variables. Extensive comparative experimental results on FG-NET facial age estimation dataset showed that the proposed method significantly improved the training efficiency, and at the same time, had higher age estimation accuracy than the state-of-the-art methods.

Key words: facial age estimation, multivariate multiple regression, partial least squares regression, label distribution learning, least square regression

中图分类号: 

  • TP391

图1

FG-NET数据库中部分人脸图像(每张图片下的数字是其对应的真实年龄)"

图2

PLS-LLD和PLSR在潜在成分数量变化下的MAE性能试验结果"

图3

PLS-LLD和CPNN-LLD在年龄标记高斯分布标准差变化下的MAE性能试验结果"

表1

4种人脸年龄估计方法的MAE和训练时间的试验结果"

方法 MAE 训练时间/s
PLS-LLD 4.55±3.29 1.65±0.54
PLSR 5.75±2.42 0.27±0.43
ⅡS-LLD 6.36±4.15 1481.32±16.87
CPNN-LLD 5.34±3.74 613.65±15.84
1 CHANG K Y , CHEN C S . A learning framework for age rank estimation based on face images with scattering transform[J]. IEEE Transactions on Image Processing, 2015, 24 (3): 785- 798.
2 FU Y , GUO G , HUANG T S . Age synthesis and estimation via faces: a survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32 (11): 1955- 1976.
3 王先梅, 梁玲燕, 王志良, 等. 人脸图像的年龄估计技术研究[J]. 中国图象图形学报, 2012, 17 (6): 603- 618.
WANG Xianmei , LIANG Lingyan , WANG Zhiliang , et al. Age estimation by facial image: a survey[J]. Journal of Image and Graphics, 2012, 17 (6): 603- 618.
4 LANITIS A , TAYLOR C J , COOTES T F . Toward automatic simulation of aging effects on face images[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24 (4): 442- 455.
doi: 10.1109/34.993553
5 LANITIS A , DRAGANOVA C , CHRISTODOULOU C . Comparing different classifiers for automatic age estimation[J]. IEEE Transactions on Systems, Man and Cybernetics, Part B(Cybernetics), 2004, 34 (1): 621- 628.
6 FU Y , HUANG T S . Human age estimation with regression on discriminative aging manifold[J]. IEEE Transactions on Multimedia, 2008, 10 (4): 578- 584.
doi: 10.1109/TMM.2008.921847
7 GUO G D , FU Y , DYER C R , et al. Image-based human age estimation by manifold learning and locally adjusted robust regression[J]. IEEE Transactions on Image Processing, 2008, 17 (7): 1178- 1188.
8 PONTES J K , BRITTO A S Jr , FOOKES C , et al. A flexible hierarchical approach for facial age estimation based on multiple features[J]. Pattern Recognition, 2016, 54, 34- 51.
doi: 10.1016/j.patcog.2015.12.003
9 GENG X , YIN C , ZHOU Z H . Facial age estimation by learning from label distributions[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35 (10): 2401- 2412.
10 GENG X . Label distribution learning[J]. IEEE Transactions on Knowledge and Data Engineering, 2016, 28 (7): 1734- 1748.
doi: 10.1109/TKDE.2016.2545658
11 王惠文. 偏最小二乘回归方法及其应用[M]. 北京: 国防工业出版社, 1999.
WANG Huiwen . Partial least-squares regression-method and application[M]. Beijing: National Defense Industry Press, 1999.
12 ABDI H . Partial least squares regression and projection on latent structure regression (PLS Regression)[J]. Wiley Interdisciplinary Reviews: Computational Statistics, 2010, 2 (1): 97- 106.
doi: 10.1002/wics.51
13 GENG X, WANG Q, XIA Y. Facial age estimation by adaptive label distribution learning[C]// 201422nd International Conference on Pattern Recognition, 24-28 Aug. 2014, Stockholm, Sweden: IEEE, 2014: 4465-4470.
14 TABACHNICK B G , FIDELL L S . Using multivariate statistics(5th ed.)[M]. Boston, MA: Allyn & Bacon/Pearson Education, 2007.
15 ROSIPAL R, KRÄMER N. Overview and recent advances in partial least squares[M]// ROSIPAL R, KRÄMER N. eds. Subspace, Latent Structure and Feature Selection. Berlin, Heidelberg: Springer Berlin Heidelberg, 2006: 34-51.
16 BOULESTEIX A L . PLS dimension reduction for classification with microarray data[J]. Statistical Applications in Genetics and Molecular Biology, 2004, 3 (1): 1- 30.
17 曾雪强, 李国正. 基于偏最小二乘降维的分类模型比较[J]. 山东大学学报(工学版), 2010, 40 (5): 41- 47.
ZENG Xueqiang , LI Guozheng . An examination of a classification model with partial least square based dimension reduction[J]. Journal of Shandong University(Engineering Science), 2010, 40 (5): 41- 47.
18 GUO G, MU G. Simultaneous dimensionality reduction and human age estimation via kernel partial least squares regression[C]//Computer Vision and Pattern Recognition. Washington, DC, USA: IEEE, 2011: 657-664.
19 COOTES T F, EDWARDS G J, TAYLOR C J. Active appearance models[C]//European Conference on Computer Vision. Heidelberg, Berlin: Springer-Verlag, 1998: 681-685.
[1] 汪嘉晨,唐向红,陆见光. 轴承故障诊断中特征选取技术[J]. 山东大学学报 (工学版), 2019, 49(2): 80-87, 95.
[2] 张红斌,邱蝶蝶,邬任重,朱涛,滑瑾,姬东鸿. 基于极端梯度提升树算法的图像属性标注[J]. 山东大学学报 (工学版), 2019, 49(2): 8-16.
[3] 侯霄雄,许新征,朱炯,郭燕燕. 基于AlexNet和集成分类器的乳腺癌计算机辅助诊断方法[J]. 山东大学学报 (工学版), 2019, 49(2): 74-79.
[4] 杨煦,陈辉,林游思,屠长河. 飞行蝙蝠标记自动提取与追踪算法[J]. 山东大学学报 (工学版), 2019, 49(2): 67-73.
[5] 胡云,张舒,李慧,佘侃侃,施珺. 基于信任网络重构的推荐算法[J]. 山东大学学报 (工学版), 2019, 49(2): 42-46.
[6] 高明霞,李经纬. 基于word2vec词模型的中文短文本分类方法[J]. 山东大学学报 (工学版), 2019, 49(2): 34-41.
[7] 李童,马然,郑鸿鹤,安平,胡翔宇. 基于视频统计特征的差错敏感度模型[J]. 山东大学学报 (工学版), 2019, 49(2): 116-121.
[8] 秦军,张远鹏,蒋亦樟,杭文龙. 多代表点自约束的模糊迁移聚类[J]. 山东大学学报 (工学版), 2019, 49(2): 107-115.
[9] 李力钊,蔡国永,潘角. 基于C-GRU的微博谣言事件检测方法[J]. 山东大学学报 (工学版), 2019, 49(2): 102-106, 115.
[10] 刘世光,王海荣,刘锦. 快速四点一致性点云粗配准算法[J]. 山东大学学报 (工学版), 2019, 49(2): 1-7.
[11] 钱春琳,张兴芳,孙丽华. 基于在线评论情感分析的改进协同过滤推荐模型[J]. 山东大学学报 (工学版), 2019, 49(1): 47-54.
[12] 周荣翔,贾修一. 中文反语识别特征分析[J]. 山东大学学报 (工学版), 2019, 49(1): 41-46.
[13] 杨思,李思童,张进东,白羽. 高速光通信激光器带宽模型改进与并行计算优化[J]. 山东大学学报 (工学版), 2019, 49(1): 17-22, 29.
[14] 范君,业巧林,业宁. 基于改进的有监督无参局部保持投影算法的人脸识别[J]. 山东大学学报 (工学版), 2019, 49(1): 10-16.
[15] 严云洋,张慧珍,刘以安,高尚兵. 基于GMM与三维LBP纹理的视频火焰检测[J]. 山东大学学报 (工学版), 2019, 49(1): 1-9.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 王素玉,艾兴,赵军,李作丽,刘增文 . 高速立铣3Cr2Mo模具钢切削力建模及预测[J]. 山东大学学报(工学版), 2006, 36(1): 1 -5 .
[2] 李梁,罗奇鸣,陈恩红. 对象级搜索中基于图的对象排序模型(英文)[J]. 山东大学学报(工学版), 2009, 39(1): 15 -21 .
[3] 李可,刘常春,李同磊 . 一种改进的最大互信息医学图像配准算法[J]. 山东大学学报(工学版), 2006, 36(2): 107 -110 .
[4] 季涛,高旭,孙同景,薛永端,徐丙垠 . 铁路10 kV自闭/贯通线路故障行波特征分析[J]. 山东大学学报(工学版), 2006, 36(2): 111 -116 .
[5] 秦通,孙丰荣*,王丽梅,王庆浩,李新彩. 基于极大圆盘引导的形状插值实现三维表面重建[J]. 山东大学学报(工学版), 2010, 40(3): 1 -5 .
[6] 刘文亮,朱维红,陈涤,张泓泉. 基于雷达图像的运动目标形态检测及跟踪技术[J]. 山东大学学报(工学版), 2010, 40(3): 31 -36 .
[7] 孙国华,吴耀华,黎伟. 消费税控制策略对供应链系统绩效的影响[J]. 山东大学学报(工学版), 2009, 39(1): 63 -68 .
[8] 孙炜伟,王玉振. 考虑饱和的发电机单机无穷大系统有限增益镇定[J]. 山东大学学报(工学版), 2009, 39(1): 69 -76 .
[9] 孙殿柱,朱昌志,李延瑞 . 散乱点云边界特征快速提取算法[J]. 山东大学学报(工学版), 2009, 39(1): 84 -86 .
[10] 杨发展1 ,艾兴1 ,赵军1 ,侯建锋2 . ZrO2含量对WC基复合材料的力学性能和微观结构的影响[J]. 山东大学学报(工学版), 2009, 39(1): 92 -95 .