• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2006, Vol. 28 ›› Issue (9): 60-63.

• 论文 • 上一篇    下一篇

一种分割虚拟人彩色冰冻切片图像的混和方法

薛志东[1,3] 隋卫平[2] 王乘[3]   

  • 出版日期:2006-09-01 发布日期:2010-05-20

  • Online:2006-09-01 Published:2010-05-20

摘要:

彩色冰冻切片图像的分割是数字虚拟人项目最为关键和基础的部分。本文提出一种新的基于支持向量机和分水岭算法相结合的混和分割方法,它利用分水岭算法修正传统支持向量方法获得的分割结果完成最终分割。实验表明,该方法可以更准确地对虚拟人彩色冰冻切片图像进行分割。

关键词: 虚拟人 支持向量机 分水岭算法 图像分割

Abstract:

The segmentation of color cryo-section images is the key and toundational step in the Virtual tauman Project. A new hybrid segmentation method based o n support vector machines and the watershed algorithm is proposed. The method segments the color cryo-section image by using the watershed algorithm to  revise the segmentation result obtained by the support vector machine. The experiment illustrates that the proposed method can segment color cryo-sectio  n images more accurately.

Key words: (virtual human, support vector machines, watershed algorithm, image segmentation)