重庆理工大学学报(自然科学) ›› 2023, Vol. 37 ›› Issue (3): 162-171.

• 信息·计算机 • 上一篇    下一篇

融合显著性检测的图像检索方法研究

田 枫,卢圆圆,刘 芳   

  1. 1.东北石油大学 计算机与信息技术学院,黑龙江 大庆 163318; 2.东北石油大学 地球科学学院,黑龙江 大庆 163318)
  • 出版日期:2023-04-26 发布日期:2023-04-26
  • 作者简介:田枫,男,博士,教授,主要从事计算机视觉、人工智能、地质资源智能信息处理研究,Email:2411318258@qq. com;通信作者 卢圆圆,女,硕士研究生,主要从事计算机视觉研究,Email:lyy1999202106@163.com。

Research on image retrieval methods based on saliency detection

  • Online:2023-04-26 Published:2023-04-26

摘要: 针对在大量图像中进行图像检索的准确度不高的问题,提出了一种显著性检测和 卷积神经网络相结合的两阶段图像检索模型 NLVG。在模型的第一阶段使用局部特征图与全 局特征图相结合的非局部深度特征模型(NLDF)进行显著性检测;在第二阶段使用 VGG16卷 积神经网络进行特征提取得到特征向量,将得到的特征向量利用相似性度量方法和建立的图像 检索库相匹配并显示与之相似的图像;使用交互式界面工具包 PyQt5设计图像检索系统实现检 索任务。使用网络爬虫技术获取图片并预处理构建数据集,对数据集上所有图像通过两阶段的 显著性检测模型进行检测得到图像特征库。实验结果表明:所提出的检索算法 map值为 0767,相较于 SpoC等算法精度有所提高,查询结果更符合预期。

关键词: 图像检索, 显著性检测, 卷积神经网络, VGG, NLDF

Abstract:  Aiming at a low accuracy of image retrieval in a large number of images, this paper proposes a two-stage image retrieval model NL-VG which combines significance detection and convolution neural network. Firstly, in the first stage of the model, a nonlocal depth feature (NLDF) model combined with the local feature map and the global feature map is used to detect the saliency. Secondly, VGG-16 convolutional neural network is used to extract feature vectors in the second stage, which are then matched with the established image retrieval database through the similarity measurement method and similar images are displayed. Finally, the interactive interface toolkit PyQt5 is used to design the image retrieval system to realize the retrieval task. In this paper, web crawler technology is used to obtain and preprocess images to construct data sets. All images on the data sets are detected by the two-stage saliency detection model to obtain the image feature database. The experimental results show that the map value of the retrieval algorithm proposed in this paper is 0.767, which is more accurate than that of SpoC and other algorithms, and the query results are more consistent with the query expectations.

中图分类号: 

  • TP39