SOS!Please help me!

[Alexey]

I can not give up metadata

I do not use metadata for K-mean

Could you please clarify?

I do not know my think right or wrong?

Grey level intensities of RGB image channels should be considered as feature vectors to which the k-mean algorithm is then applied to segment your picture. So you had better care about how to represent your binary data (stored in a blob field) as feature vectors, then about how to use Oracle GeoRaster model.

[Thoa]

Hi Alexey!
Thanks for your help!
I understand the K-mean algorithm, let me display some following problems:
1> I loaded all cellvalue of my Image into Array of C# (include Sizeof Image and R,G,B at each pixel). I think from this, I can use K-mean, if need gray level at each pixel, GeoRaster has support this (sdo_geor.GrayScale) so that I can get Gray Scale for all pixels in my Image. It is right or wrong?
2> Now, I just have sizeofImage and value R,G,B of them. It means I have color value of my Image not an Image in a blob. is it enough for me to use K-mean, what anything else I need. I am very worried so please help me and tell me what I draw is right or wrong?
Thanks so much!
THoa.

[Alexey]

Which characteristics of the image you would pass to K-mean algorithm is up to you. It can be intensity of each pixel, its gradient, etc. We cannot give you advices on using segmentation algorithms because this exceeds goals of our support.

[Thoa]

Hi Alexey!
Thank for your soon reply!
I just would like to know if I have array include all pixels and color values of them then I can begin with image processing?
Please let me know about that!
Thanks!
Thoa.

[Alexey]

I think so. You should have array with values of colourness to pass to segmentation algorithm.

[Thoa]

Hi Alexey!
Thanks so much for your help!
Your advice make me more confident by myself to implement segment algorithm. Thanks so much! If I have any problem, I always would like your help, Could you help me!
Have nice weekend!
Thoa!

[Alexey]

You are always welcome.