Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (27): 191-192.DOI: 10.3778/j.issn.1002-8331.2009.27.058

• 图形、图像、模式识别 • Previous Articles     Next Articles

Image local segmentation based on multi-fractal theory

XIA Zheng-wei,ZHAO Jian,PU Xiao-qin,GUO Xiu-mei   

  1. School of Information Science and Technology,Northwest University,Xi’an 710069,China
  • Received:2008-05-19 Revised:2008-08-15 Online:2009-09-21 Published:2009-09-21
  • Contact: XIA Zheng-wei

多重分形应用于图像局部分割的研究

夏政伟,赵 健,蒲小勤,郭秀梅   

  1. 西北大学 信息科学与技术学院,西安 710069
  • 通讯作者: 夏政伟

Abstract: In view of the existing different image segmentation methods,this paper uses a partial image segmentation method based on the multi-fractal Hölder exponents and the multi-fractal singular spectrum,the method can maintain and strengthen the image of local features,and extract some interested sense of image.Experiments show that this method achieves a better division,and gets an ideal image of segmentation results.

Key words: multi-fractal theory, Hö, lder exponents, multi-spectral singular spectrum, image local segmentation

摘要: 在现有对图像分割方法研究的基础上,采用基于多重分形的Hölder指数和多重分形奇异谱联合考虑的图像局部分割方法。方法具有保持和加强图像局部特征,对图像中某些敏感部分进行边缘提取,较好实现图像分割的功能。实验表明该方法能够达到较好的图像分割效果,具有实时性、有效性和可实现性。

关键词: 多重分形, Hö, lder指数, 多重分形奇异谱, 图像局部分割

CLC Number: