首页    期刊浏览 2024年09月29日 星期日
登录注册

文章基本信息

  • 标题:Spectral-based Shadow Detection for Single Image
  • 本地全文:下载
  • 作者:Xiaoyan Xu ; Xiaoming Liu
  • 期刊名称:International Journal of Hybrid Information Technology
  • 印刷版ISSN:1738-9968
  • 出版年度:2015
  • 卷号:8
  • 期号:4
  • 页码:157-164
  • DOI:10.14257/ijhit.2015.8.4.18
  • 出版社:SERSC
  • 摘要:Shadow detection for single image is difficult but has wide applications. This paper proposes a novel and fast shadow detection method based on the Tricolor Attenuation Model [1]. In this study, we analyze the spectral property of outdoor light sources to estimate the parameters of TAM. Then our shadow detection method is proposed by integrate the TAM feature and intensity information. Our method can extract shadows for a single and uncalibrated image without any prior knowledge, even for complex scenes. Experimental results show the effectiveness of the proposed method.
  • 关键词:Shadow detection; Tricolor attenuation model
国家哲学社会科学文献中心版权所有