+高级检索
用于交通标志检测的窗口大小聚类残差SSD模型
DOI:
作者:
作者单位:

长安大学信息工程学院

作者简介:

通讯作者:

基金项目:

国家自然科学基金资助项目(61201406, 61572083);中央高校基本科研业务费专项资金资助项目(310824162022)


A residual SSD model based on window size clustering for traffic sign detection
Author:
Affiliation:

1.School of Information Engineering, Chang'2.'3.an University;4.School of Information Engineering, Chang'an University

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
    摘要:

    SSD通常被认为适合于求解小目标图像检测问题,但在特征表征和检测效率两方面还存在改进空间.本文提出一种聚类残差SSD模型,一方面将原始SSD模型中的VGG16基础网络替换为更深的ResNet50残差网络,以改善特征表征能力.另一方面采用K-均值聚类算法取代盲目搜索机制,确定SSD中默认窗口的大小,以改善检测效率.针对德国交通标志检测数据集,模型获得了97.1% mAP和每幅图像0.07s的检测速度.针对中国交通标志数据集,模型获得89.7% mAP和每幅图像0.08s的检测速度.与原始SSD模型比较,所提模型的检测性能得到改善.

    Abstract:

    Single Shot MultiBox Detector (SSD) is generally considered to be suitable for solving small target detection in images, its performance on feature extraction and detection efficiency is still required to be improved, however. A clustering residual SSD model is proposed in this paper. On one hand, in order to improve the feature extraction quality the basic network VGG16 which is consisted of the orginal SSD model is replaced with a deeper residual network ResNet50. On the other hand, in order to improve the detection efficiency, K-means algorithm other than the blind search mechanism used in the original SSD model is exploited to find and determine the assignments of the sizes of default windows. For German traffic sign detection dataset, it obtains 97.1% mAP in detection accuracy and 0.07s per image in detection efficiency. For Chinese traffic sign dataset, it obtains 89.7% mAP in detection accuracy and 0.08s per image in detection efficiency. Compared with the original SSD model, the proposed model obtains improved detection performance.

    参考文献
    相似文献
    引证文献
文章指标
  • PDF下载次数:
  • HTML阅读次数:
  • 摘要点击次数:
  • 引用次数:
引用本文
历史
  • 收稿日期: 2018-11-21
  • 最后修改日期: 2019-01-04
  • 录用日期: 2019-01-23
  • 在线发布日期:
  • 出版日期:
作者稿件一经被我刊录用,如无特别声明,即视作同意授予我刊论文整体的全部复制传播的权利,包括但不限于复制权、发行权、信息网络传播权、广播权、表演权、翻译权、汇编权、改编权等著作使用权转让给我刊,我刊有权根据工作需要,允许合作的数据库、新媒体平台及其他数字平台进行数字传播和国际传播等。特此声明。
关闭