ASIA unversity:Item 310904400/12745
English  |  正體中文  |  简体中文  |  全文笔数/总笔数 : 90069/105176 (86%)
造访人次 : 6421367      在线人数 : 595
RC Version 6.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
搜寻范围 查询小技巧:
  • 您可在西文检索词汇前后加上"双引号",以获取较精准的检索结果
  • 若欲以作者姓名搜寻,建议至进阶搜寻限定作者字段,可获得较完整数据
  • 进阶搜寻
    ASIA unversity > 資訊學院 > 資訊工程學系 > 博碩士論文 >  Item 310904400/12745


    题名: Evaluation of Color Descriptors for Image Matching Under Changing Illumination Conditions
    作者: Li, Yi-Pin
    贡献者: Department of Computer Science and Information Engineering
    Ng, Hui-Fuang
    关键词: Illumination Changes;Image Matching;LBP;SIFT;Color Descriptors
    日期: 2012
    上传时间: 2012-11-18 17:00:56 (UTC+8)
    出版者: Asia University
    摘要: Color descriptors have been used extensively and successfully in many computer vision applications. However, object colors are sensitive to changes in illumination conditions such as lighting geometry and illumination color. Changes in the illumination of a scene can greatly affect the performance of image matching and object recognition if the color descriptors used are not robust to these changes. This study examined and compared the invariance properties of various state-of-the-art color descriptors and evaluated their performance on image matching under changing illumination conditions. We evaluated the performance of the color descriptors on image matching under changing illumination conditions using the Amsterdam Library of Object Images (ALOI) database, which is an image database of colored objects taken under various imaging conditions. Experimental results indicate that SIFT-based descriptors and LBP-based descriptors are less sensitive to changes in illumination conditions. Comparing to other color descriptors, the SIFT-Opponent descriptor has the best overall performance. The Opponent color descriptor, on the other hand, has the lowest correct match rates.
    显示于类别:[資訊工程學系] 博碩士論文


    档案 描述 大小格式浏览次数


    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - 回馈