欢迎访问《图学学报》 分享到：
图学学报

图学学报 ›› 2021, Vol. 42 ›› Issue (1): 1-7.DOI: 10.11996/JG.j.2095-302X.2021010001

• 图像处理与计算机视觉 • 上一篇    下一篇

基于深度学习与有向无环图 SVM 的局部调整年龄估计

  

  1. (1. 成都工业学院教务处教学建设与教学质量管理科，四川 成都 611730； 2. 成都工业学院计算机工程学院，四川 成都 611730)
  • 出版日期:2021-02-28 发布日期:2021-01-29
  • 基金资助:
    四川省教育厅《省级教育体制机制改革试点项目》(G5-08) 

Locally adjusted age estimation based on deep learning and directed acyclic graph SVM  

  1. (1. Teaching Construction and Teaching Quality Management Section, Department of Education, Chengdu Technological University, Chengdu Sichuan 611730, China; 2. School of Computer Engineering, Chengdu Technological University, Chengdu Sichuan 611730, China)
  • Online:2021-02-28 Published:2021-01-29
  • Supported by:
    Provincial Education System and Mechanism Reform Pilot Project of Sichuan Provincial Education Department (G5-08) 

PDF (PC)

334

摘要： 为了进一步从人脸图像中提高年龄估计的精度，提出一种基于深度学习与有向无环图支持向量 机(SVM)的局部调整年龄估计算法。在训练阶段，首先将经过 VGGFace2 数据集预训练的 SE-ResNet-50 网络 进行微调，并在收敛时提取全连接层，将其首尾相连形成的向量作为表征并训练得到多个 one-versus-one SVM； 在测试阶段，先将待估计人脸图像送入 SE-ResNet-50 以得到一个较为粗略的年龄估计值，然后设定具体邻域， 最后将训练而成的 SVM 组合为一个有向无环图 SVM 并以全局估计值为中心进行精准的年龄估计。为了表明 算法的普适性，在不同种族的 MORPH 和 AFAD 图像集中进行了实验，结果验证了算法的有效性。

关键词: 年龄估计, 深度学习, 有向无环图支持向量机, 局部调整

Abstract:  In order to further enhance the accuracy of age estimation, we proposed a locally adjusted age estimation algorithm based on deep learning and directed acyclic graph-support vector machine (SVM). In the training phase, the SE-ResNet-50 network, pre-trained on the VGGFace2 data set, was first fine-tuned. When it converged, the fully connected layer was extracted, and the vector formed by its end-to-end connection was employed as a representation and further trained multiple one-versus-one SVM. In the testing phase, we first sent the face image into SE-ResNet-50 to obtain a rough age result, then set the specific neighborhood, finally integrated the trained SVM into a directed acyclic graph SVM, and conducted accurate age estimation centering on the global estimation value. In order to show the universality of the algorithm, the results of experiments undertaken in MORPH and AFAD datasets of different races can verify the effectiveness of the algorithm. 

Key words: age estimation, deep learning, directed acyclic graph support vector machine, local adjustment  

中图分类号: