计算机应用 ›› 2012, Vol. 32 ›› Issue (06): 1526-1528.DOI: 10.3724/SP.J.1087.2012.01526

• 图形图像技术 • 上一篇    下一篇

基于分解的三维Otsu图像分割快速算法

龚劬,倪麟,唐萍峰,王菲菲   

  1. 重庆大学 数学与统计学院,重庆 401331
  • 收稿日期:2011-11-23 修回日期:2012-01-14 发布日期:2012-06-04 出版日期:2012-06-01
  • 通讯作者: 龚劬
  • 作者简介:龚劬(1963-),女,四川威远人,教授,博士,主要研究方向:图像处理、小波分析;〓倪麟(1987-),男,四川威远人,硕士研究生,主要研究方向:图像处理;〓唐萍峰(1987-),男,四川广安人,硕士研究生,主要研究方向:模式识别、图像处理;〓王菲菲(1987-),女,河南濮阳人,硕士研究生,主要研究方向:图像处理。

Fast three-dimensional Otsu image segmentation algorithm based on decomposition

GONG Qu,NI Lin,TANG Ping-feng,WANG Fei-fei   

  1. College of Mathematics and Statistics, Chongqing University, Chongqing 401331,China
  • Received:2011-11-23 Revised:2012-01-14 Online:2012-06-04 Published:2012-06-01
  • Contact: GONG Qu

摘要: 摘要:针对三维Otsu图像分割算法计算复杂度高,运算量大的问题,本文提出一种基于分解的三维Otsu图像分割快速算法。该算法,首先将三维Otsu分解为三个一维Otsu;然后,在分析一维Otsu的基础上,结合类间距离和类内距离,提出一种新的阈值识别函数设计算法,并给出了快速实现方法。实验结果表明,该算法不仅可以取得较好的分割效果,而且计算量较小,比三维Otsu阈值分割递推算法快1400倍左右。

关键词: 图像分割, Otsu法, 类间距离, 类内距离, 阈值识别函数

Abstract: Abstract: Aiming at the weakness of the computational complexity and huge calculation of the three-dimensional Otsu, a fast three-dimensional Otsu image segmentation algorithm based on decomposition was presented in this paper. Firstly, the original three-dimensional Otsu algorithm was decomposed into three one-dimensional Otsu algorithms. Then, based on the one-dimensional Otsu algorithm, a novel algorithm with a new threshold recognition function was proposed, which combines between-class distance with within-class distance, and the fast realization method was also presented. The experimental results show that the proposed algorithm does not only get satisfactory segmentation result, but also improves the calculation speed, which spends 1400 times less than the recursive algorithm for the three-dimensional Otsu method.

Key words: image segmentation, Otsu algorithm, between-class distance, within-class distance, threshold recognition function

中图分类号: