English

应用科学学报 ›› 1998, Vol. 16 ›› Issue (3): 372-374.

• 论文 • 上一篇    下一篇

图象自相似性测度算法的研究比较

董远1, 吴小滔2, 胡光锐1   

  1. 1. 上海交通大学;
    2. 上海第二工业大学
  • 收稿日期:1997-06-12 修回日期:1997-10-09 出版日期:1998-09-30 发布日期:1998-09-30
  • 作者简介:董远:博士,上海交通大学电子工程系,上海 200030
  • 基金资助:
    国家自然科学基金

Comparison of the Algorithms on Image Similarity Measure

DONG YUAN1, WU XIAOTAO2, HU GUANGRUI1   

  1. 1. Shanghai Jiaotong University, Electronic Engineering Department, shanghai 200030;
    2. Shanghai Secondary Polytechnic University, shanghai 200041
  • Received:1997-06-12 Revised:1997-10-09 Online:1998-09-30 Published:1998-09-30

PDF

122

摘要： 图象的自相似性测度可作为图象识别和图象压缩的重要特征参量.分形几何图形具有自相似性和递归性,分形维数自然成为有效的自相似性测度.基于目前常用的有效的计算图象分形维数算法,其中包括"毯子"法、"盒子"法、"分数布朗随机场模型"法和各种改进算法,对各种方法的计算量、适用范围进行了研究比较.

关键词: 分形, 分形维数, 纹理

Abstract: The similarity measure of image could be a useful character in image recognition and compression. Fracal pictures possess similarity and recursion, and fractal dimension naturally becomes a effective measure of similarity. The paper mainly discusses the complexity and applicability of the effective fractal dimension algorithms interested in these years, including blanket method, boxing method, fraction Brown random model method and other methods improved.

Key words: fractal dimension, texture, fractal