摘要
提出了一种基于同组的图像分割算法。首先Fisher判据对初始聚类中心自动确定出一个同组;然后,再用置信度来判定这一同组是否有效,从而消除过分割,实现图像的精确分割。最后,给出了上述算法的计算机实现方法,并进行了计算机仿真研究。
This paper proposes an image segmentation method based on peer group technology. Firstly, the Fisher discrimination algorithm is used to get a peer group corresponding to an initial cluster automatically. Secondly, a confidence magnitude is used to testify the validation of the peer group so that the over-segmentation can be eliminated and the accurate image segmentation can be implemented. At last, the computer realization steps of the algorithm mentioned above are presented, and computer simulation researches are performed on the algorithm.
出处
《科技信息》
2010年第5期61-61,53,共2页
Science & Technology Information
关键词
图像分割
Fisher判据
同组
Image segmentation
Fisher discrimination
Peer group
