How can the center be determined by the Kmeans method?

Question

Ansam Nazar 2020-6-24

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/553858-how-can-the-center-be-determined-by-the-kmeans-method

回答： Aakash Mehta 2020-6-24

Hello
I am trying to use the Kmeans algorithm of the classification process to separate the diseases in the rays from the normal, but every time the colors resulting from the Kmeans process change. I want to know how I can confirm the colors resulting from each implementation so that each color is specific to a particular class in the xray.
indexes = kmeans(grayImage(:), numberOfClasses);

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

KSSV 2020-6-24

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/553858-how-can-the-center-be-determined-by-the-kmeans-method#answer_456382

在 MATLAB Online 中打开

[idx,C,D] = kmeans(grayImge(:),numberOfclasses) ;

C is your cneter.

Read about kmeans. You can your center along with the function. The output is random..the classe indices change for every run.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Ansam Nazar 2020-6-24

Thank you so much

But I am looking for a way to stabilize the resulting colors and cancel random with each implementation

请先登录，再进行评论。

Answer 2

Aakash Mehta 2020-6-24

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/553858-how-can-the-center-be-determined-by-the-kmeans-method#answer_456433

Due to the random starting value of the kmeans algorithm you are getting the different results with each implementation.

In order to get the results closer in each implementation,

Use the 'Start' property of kmeans algorithm. here, you can speecify the start points for your algorithm.so, each time kmeans algorithm starts from those points.
Also increase the no of iterations using the 'MaxIter' property.

For more help regarding name-value pair, please go to kmeans name-value pair.