How do I classify points in an image into groups?

Hello to you all

I have an image that contains points and I want to combine these points based on the distance between them and finally surround each group;

Thank you for your help, and here attached an example of the image

Answer 1

[Salaheddin Hosseinzadeh]

Hi Maa Kari,

You can shrink the spots to a single point using bwmorph().

Then you may classify them according to their X and Y position, or you can use a perceptron for classification.

Good luck ;)

Answer 2

[Maa Kari]

Hi Salaheddin Hosseinzadeh thanks for yur answer;

Yes i have a vector (n x 2 ) that contain the X and Y posistion of my points , but i dont have any idea how to do , can you please clarify

Answer 3

[Image Analyst]

It sounds like you don't have an image but actually just a list of x,y coordinates. Do you have the Statistics and Machine Learning Toolbox? There are functions in there for clustering/classification. For example, you could use kmeans().

[Image Analyst]

Maa's "Answer" moved here since it's a reply to me and not an "Answer" to the original question:

yes you are right, Assume that i have a list of x,y coordinates , but if i use a kmeans() i must give the (k) number of clusters but i dont konw exactly the number of clusters it changes from an image to an other one,

thanks

[Image Analyst]

Then use one of the other unsupervised methods. What are the two feature along the x and y axis? What image property do they represent? Would you care to show the image?

[Maa Kari]

there are no feature for the X and Y axes,the image is above it's just a show of points using plot(), my problem is that I have a matrix that contains coordinates of the point, and I want to group them according to some criterion and have sub-matrix of the coordinates of the points of each group , can you please tell me some unsupervised methods Thanks

[Image Analyst]

The Statistics and Machine Learning Toolbox provides supervised and unsupervised machine learning algorithms, including support vector machines (SVMs), boosted and bagged decision trees, k-nearest neighbor, k-means, k-medoids, hierarchical clustering, Gaussian mixture models, and hidden Markov models.

See Examples and how-tos here: http://www.mathworks.com/discovery/unsupervised-learning.html