Sir, I have a question regarding co-occurrence matrix: What would be the size of co-occurence matrix if have an image region (5x5) having pixel values from 164 to 168? Could you clarify please?
As long as it is a 8-bit image, the matrix is always 256x256, whatever the size of the region you take. In your case, the matrix would be really sparse. I guess you may reduce the matrix if the image only has a tiny range of pixel value.
You just slide a 3x3 kernel and calculate the LBP value for the pixel in the middle, just like convolution. In your case, you will get a 2x4 feature map.
Thank you very much! Great explanation.
sir could you suggest a book for reading more about texture and how I start learning from it to extend some thing like GLCM for my studying?
Thank you sir...
Sir, I have a question regarding co-occurrence matrix: What would be the size of co-occurence matrix if have an image region (5x5) having pixel values from 164 to 168? Could you clarify please?
As long as it is a 8-bit image, the matrix is always 256x256, whatever the size of the region you take. In your case, the matrix would be really sparse. I guess you may reduce the matrix if the image only has a tiny range of pixel value.
Hello Sir, How to find LBP when there is matrix of size 4 by 6
You just slide a 3x3 kernel and calculate the LBP value for the pixel in the middle, just like convolution. In your case, you will get a 2x4 feature map.