Journal of Computer Applications ›› 2023, Vol. 43 ›› Issue (5): 1578-1583.DOI: 10.11772/j.issn.1001-9081.2022040606

• Multimedia computing and computer simulation • Previous Articles    

Multi-task age estimation method based on multi-peak label distribution learning

Jianhui HE, Chunlong HU(), Xin SHU   

  1. School of Computer Science and Engineering,Jiangsu University of Science and Technology,Zhenjiang Jiangsu 212100,China
  • Received:2022-04-29 Revised:2022-07-06 Accepted:2022-07-07 Online:2023-05-08 Published:2023-05-10
  • Contact: Chunlong HU
  • About author:HE Jianhui, born in 1998, M. S. candidate. His research interests include computer vision, age estimation.
    HU Chunlong, born in 1985, Ph. D., associate professor. His research interests include biometric identification, deep learning.
    SHU Xin, born in 1979, Ph. D., associate professor. His research interests include pattern recognition, computer vision.
  • Supported by:
    Jiangsu University of Science and Technology Scientific Research Fund(1132921402)

基于多峰标签分布学习的多任务年龄估计方法

何建辉, 胡春龙(), 束鑫   

  1. 江苏科技大学 计算机学院,江苏 镇江 212100
  • 通讯作者: 胡春龙
  • 作者简介:何建辉(1998—),男,湖南永州人,硕士研究生,主要研究方向:计算机视觉、年龄估计
    胡春龙(1985—),男,江苏盐城人,副教授,博士,主要研究方向:生物特征识别、深度学习 huchunlong@just.edu.cn
    束鑫(1979—),男,江苏丹阳人,副教授,博士,CCF会员,主要研究方向:模式识别、计算机视觉。
  • 基金资助:
    江苏科技大学科学研究基金资助项目(1132921402)

Abstract:

Considering the difficulty of extracting label ordinal information and inter-class correlation in facial age estimation, a Multi-Peak Distribution (MPD) age coding was proposed, and a multi-task age estimation method MPDNet (MPD Network) was constructed based on the proposed age coding. Firstly, in order to extract correlation information among age labels and construct aging trend stages, the age labels were transformed into age distributions by using MPD. Then, a lightweight network was used for multi-stage feature extraction, and Label Distribution Learning (LDL) and regression learning were performed on the extracted features respectively. Finally, the outputs of the two learning tasks were shared and optimized with each other by back-propagation during the training process, thereby avoiding the error propagation caused by the direct regression of distribution results in traditional label distribution learning. Experimental results on MORPH Ⅱ dataset show that, the Mean Absolute Error (MAE) of MPDNet reaches 2.67, which is similar to that of the methods such as DEX (Deep EXpectation) and RankingCNN (Ranking Convolutional Neural Network) built by VGGNets (Visual Geometry Group Networks), while the parameters of MPDNet are only 1/788.6 of those of VGGNets. Meanwhile, MPDNet outperforms lightweight methods such as C3AE and SSR-Net (Soft Stagewise Regression Network). MPDNet can better utilize the rich correlation information among age labels to extract more discriminative age features and improve the prediction accuracy of age estimation tasks.

Key words: age estimation, age coding, Label Distribution Learning (LDL), multi-task learning, Convolutional Neural Network (CNN)

摘要:

针对面部年龄估计中标签序数信息和类间相关性提取难的问题,提出一种多峰分布(MPD)年龄编码,并基于该年龄编码构建了一个多任务年龄估计方法MPDNet(MPD Network)。首先,利用MPD将年龄标签转化为年龄分布,以提取年龄标签间的相关信息,构建年龄老化趋势的阶段性;然后,采用一个轻量级网络进行多阶段的特征提取,并对提取的特征分别进行标签分布学习(LDL)和回归学习;最后,共享两个学习任务的输出,并在训练过程中通过反向传播互相优化,避免传统标签分布学习中对分布结果直接进行回归导致的误差传播。在MORPH Ⅱ数据集上的实验结果表明,MPDNet的平均绝对误差(MAE)达到2.67,与基于VGGNets (Visual Geometry Group Networks)构建的DEX(Deep EXpectation)、RankingCNN (Ranking Convolutional Neural Network)等方法相当,而参数仅为VGGNets的1/788.6;而且MPDNet也优于同体量的C3AE(extremely Compact yet efficient Cascade Context-based Age Estimation model)、SSR-Net (Soft Stagewise Regression Network)等方法。MPDNet能够较好地利用年龄标签间丰富的相关信息来提取更具判别力的年龄特征,提高年龄估计任务的预测精度。

关键词: 年龄估计, 年龄编码, 标签分布学习, 多任务学习, 卷积神经网络

CLC Number: