clustering rgb image block

Question

1.00/5 (1 vote)

See more:

i have a rgb image and i split into multi blocks, i want to use fuzzy c-mean algorithm for clustering of image blocks.

please help me how to cluster them when each data point such as an image block

thanks you so much.

Posted 31-Dec-14 16:08pm

Leo Phạm

Add a Solution

Comments

Sergey Alexandrovich Kryukov 31-Dec-14 22:16pm

Not really a question. Remember, on this forum, we are not supposed to just help; we help only by answering questions.
—SA

1 solution

Add a Solution

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Sergey Alexandrovich Kryukov · Answer 1 · 2014-12-31T16:20:00

Solution 1

Help with what? Fuzzy clustering in general is explained, for example, here: http://en.wikipedia.org/wiki/Fuzzy_clustering[^].
And k-means clustering — here: http://en.wikipedia.org/wiki/K-means_clustering[^].

Understanding requires some mathematical background, not very complicated, but it should include the understanding of fuzzy set mathematics:
http://en.wikipedia.org/wiki/Fuzzy_set[^],
http://en.wikipedia.org/wiki/Fuzzy_logic[^].

Algorithms are also explained; and even the extensive list of most important implementations, mostly open-source, is provided.

—SA

Posted 31-Dec-14 16:20pm

Sergey Alexandrovich Kryukov

Updated 31-Dec-14 16:21pm

v2

Comments

Leo Phạm 31-Dec-14 22:40pm

thanks SA, I already know how to cluster with each data point is just one pixel , my problem is how to cluster with each data point is an image block including many pixels .
specifically objective function and the distance formulated between data points?

Sergey Alexandrovich Kryukov 31-Dec-14 22:50pm

I don't understand what does it mean "data point is just one pixel" or "image block including many pixels". The problem is very simple: you have to define it. How? All you have to do it to mathematically define norm in a normed vector space. In other words, you have a vector space and need to give some definition for the distance between each two vectors. It's up to you how you define it, as soon as the definition of the vector space remains valid:
http://en.wikipedia.org/wiki/Normed_vector_space.

Then you apply exact same algorithm for clustering. You can even make the method defining the norm a parameter of your function performing clustering.

—SA

clustering rgb image block

1 solution

Solution 1

Add your solution here

Preview 0

clustering rgb image block

1 solution

Solution 1

Add your solution here

Preview 0

Existing Members

...or Join us