0

I need to:

  1. open an image
  2. divide it into a N x N grid
  3. then divide the Hue, Saturation and Value channels into bands of equal size
  4. then do various computations with them (but what matters is how to divide the image and then divide the channels).

Is there a way to do this without any additional frameworks/libraries (in C#)?

This is going to be a simple console application, which generates a feature vector from images, that will be used for training my SVM later.

crthompson
  • 15,653
  • 6
  • 58
  • 80
user3208638
  • 1
  • 1
  • What doe you mean by "divide it into an NxN grid"? Are you talking about taking the average of all channels in a certain group of pixels? Also what do you mean by "then divide the Hue, Saturation and Value channels into bands of equal size"? Are you referring to a type if histogram normalization for HSV? Or, are you just wanting to computer HSV for each pixel in the image, then perform computations on the image? – Matt Johnson Jan 19 '14 at 05:20
  • Thanks for the response! I already figured out most of the things. My idea was to create a boolean feature vector for the SVM, where I check if there is a single pixel from a given square (from the grid) that belongs to a certain HSV (or RGB) range. I am sorry, my english is not that good, so I probably can't explain really well, but thanks for your reply! – user3208638 Jan 20 '14 at 13:43

0 Answers0