Journal of South China University of Technology (Natural Science Edition) ›› 2010, Vol. 38 ›› Issue (11): 128-134.doi: 10.3969/j.issn.1000-565X.2010.11.023

• Computer Science & Technology • Previous Articles     Next Articles

Data Field-Based Method for Image Segmentation with Two-Dimension Threshold

Wu Tao1  Qin Kun2  Ou Lei-hai2  Du Yi3   

  1. 1.State Key Laboratory of Software Engineering,Wuhan University,Wuhan 430072,Hubei,China;2.School of Remote Sensing and Information Engineering,Wuhan University,Wuhan 430079,Hubei,China;3.Communication Network Technology Management Center,Beijing 100840,China
  • Received:2010-03-24 Revised:2010-05-23 Online:2010-11-25 Published:2010-11-25
  • Contact: 吴涛（1980-），男，博士生，讲师，主要从事智能图像处理、不确定性人工智能研究． E-mail:taowu0706@gmail．com
  • About author:吴涛（1980-），男，博士生，讲师，主要从事智能图像处理、不确定性人工智能研究．
  • Supported by:

    国家“973”计划项目（2007CB311003）; 国家自然科学基金资助项目（60875007）; 湛江市科技攻关计划项目（2009064）

PDF

926

Abstract:

In order to correctly select the optimal threshold for image segmentation,a novel method of image segmentation based on data field is proposed.The method maps the image from grayscale space to the appropriate potential space in data field,and measures the interactions of the elements in the two-dimension histogram by taking the frequency of two-dimension gray histogram as the mass of data field,thus generating a three-dimension data field.Then,by employing the potential center elimination and combination,the optimal threshold is determined and good segmentation result is obtained without significantly increasing the time complexity.It is indicated by the experiments for standard image datasets and some noisy images that,as an alternative to OTSU,the proposed me-thod is reasonable and effective with certain noise resistance.

Key words: image segmentation, threshold, two-dimension histogram, data field, OTSU, complexity