一种基于Rb-K平均带宽设定的点带宽可变的Mean Shift算法 |
| |
作者单位: | 云南大学计算机科学与工程系 云南昆明650091 |
| |
摘 要: | Mean shift是一个用在图像滤波、图像分割中的迭代过程.传统的Mean shift在迭代过程中每一个点的带宽是固定不变的,现有的Mean shift没有考虑迭代点带宽可变的情况.提出一种新的迭代点带宽可变的Mean shift,即每一个点在迭代过程中带宽是变化的.对传统方法和迭代点带宽可变方法进行比较,并对迭代点带宽可变方法进行评价,迭代点带宽可变方法在细节处理方面有好的效果.
|
关 键 词: | 聚类 无监督学习 Mean shift 带宽 模式点 |
A new bandwidth varivable mean shift based on Rb-K means |
| |
Authors: | GUO Jin-jiang WU Yi-pu ZHANG Xue-jie |
| |
Abstract: | Mean shift is an effective iterative procedure that uses in edge preserving filtering and image segmentation. Existing mean shift in the iterative process each point bandwidth is fixed, it has not considered the iteration bandwidth variable situation. A new iterative point bandwidth varivable mean shift was proposed, namely each point in the iterative process the bandwidth is changed. There are the comparison to the traditional method and the iteration bandwidth variable method and the appraisal of the iteration bandwidth variable method. The iterative point bandwidth varivable mean shift has good effect in the details of image processing. |
| |
Keywords: | clustering unsupervised learning mean shift bandwidth mode |
本文献已被 CNKI 等数据库收录! |