How to use kmeans for an image

Question

Mathew Cherian Carathedathu 2011-12-19

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/24341-how-to-use-kmeans-for-an-image

Hi..

I need to apply kmeans clustering for a grayscale image (the clustering should be based on the value of the pixels and not the distance). I tried to apply the inbuilt function to do that but the documentation says I will get only a n-by-1 vector as the output. How do I then get the cluster values for each of the pixels?

Thank you,

Mathew

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Image Analyst 2011-12-19

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/24341-how-to-use-kmeans-for-an-image#answer_31951

Here's an example that you should be able to modify from color to grayscale rather easily: http://www.mathworks.com/products/demos/image/color_seg_k/ipexhistology.html

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

Answer 2

Walter Roberson 2011-12-19

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/24341-how-to-use-kmeans-for-an-image#answer_31950

What do you mean by "the cluster values" ? kmeans() returns the cluster number for each point (each row is considered a point) as the first output. That kmeans() cannot return the cluster probabilities, but the four-output version will return the distance from each point to each centroid.