计算机应用 ›› 2018, Vol. 38 ›› Issue (8): 2198-2204.DOI: 10.11772/j.issn.1001-9081.2018020301

• 人工智能 • 上一篇    下一篇

基于深度模型迁移的细粒度图像分类方法

刘尚旺1,2, 郜翔1,2   

  1. 1. 河南师范大学 计算机与信息工程学院, 河南 新乡 453007;
    2. "智慧商务与物联网技术"河南省工程实验室(河南师范大学), 河南 新乡 453007
  • 收稿日期:2018-02-01 修回日期:2018-03-26 出版日期:2018-08-10 发布日期:2018-08-11
  • 通讯作者: 刘尚旺
  • 作者简介:刘尚旺(1973-),男,河南新乡人,副教授,博士,CCF会员,主要研究方向:生物图像处理、计算机视觉;郜翔(1992-),男,河南新乡人,硕士研究生,主要研究方向:计算机视觉、深度学习。
  • 基金资助:
    国家自然科学基金资助项目(U1304607);河南省高等学校重点科研项目(15A520080);河南师范大学博士科研启动基金资助项目(qd12138)。

Fine-grained image classification method based on deep model transfer

LIU Shangwang1,2, GAO Xiang1,2   

  1. 1. College of Computer and Information Engineering, Henan Normal University, Xinxiang Henan 453007, China;
    2. Henan Engineering Laboratory of Intelligence Business and Internet of Things(Henan Normal University), Xinxiang Henan 453007, China
  • Received:2018-02-01 Revised:2018-03-26 Online:2018-08-10 Published:2018-08-11
  • Supported by:
    This work is partially supported by the National Natural Science Foundation of China (U1304607), the Key Scientific Research Project of Higher School of Henan Province (15A520080), the Dr. Startup Project of Henan Normal University (qd12138).

摘要: 针对细粒度图像分类方法中存在模型复杂度较高、难以利用较深模型等问题,提出深度模型迁移(DMT)分类方法。首先,在粗粒度图像数据集上进行深度模型预训练;然后,使用细粒度图像数据集对预训练模型logits层进行不确切监督学习,使其特征分布向新数据集特征分布方向迁移;最后,将迁移模型导出,在对应的测试集上进行测试。实验结果表明,在STANFORD DOGS、CUB-200-2011、OXFORD FLOWER-102细粒度图像数据集上,DMT分类方法的分类准确率分别达到72.23%、73.33%和96.27%,验证了深度模型迁移方法在细粒度图像分类领域的有效性。

关键词: 深度模型, 迁移学习, 细粒度图像分类, 不确切监督学习, 特征分布

Abstract: To solve the problems of fine-grained image classification methods, such as highly complex methods and difficulty of using deeper models, a Deep Model Transfer (DMT) method was proposed. Firstly, the deep model was pre-trained on the coarse-grained image dataset. Secondly, the pre-trained deep model classification layer was trained based on inexact supervised learning by using fine-grained image dataset and transferred to the feature distribution direction of the novel dataset. Finally, the trained model was exported and tested on the corresponding test sets. The experimental results show that the classification accuracy rates on the STANFORD DOGS, CUB-200-2011 and OXFORD FLOWER-102 fine-grained image datasets are 72.23%, 73.33%, and 96.27%, respectively, which proves the effectiveness of DMT method on fine-grained image classification.

Key words: deep model, transfer learning, fine-grained image classification, inexact supervised learning, feature distribution

中图分类号: