how to train svm using glcm features

i am doing project on localizing non-uniform deblurring in an image.i've used 6 sliding windows(4*4 , 8*8 ,16*16 ,32*32 ,64*64 ,128*128) on the image and glcm features are extracted from the 1st and second order derivative of the image on each sliding winow.now i need to train them using svm so that authentic regions will be labeled as 0 and deblurred region as 1. but i am not understanding that using glcm features how can i train them.Could anyone help me

Answer 1

[Rajani Mishra]

GLCM matrix easily helps to extract the texture features.

To calculate different statistics (Example: Contrast, Correlation) from “glcm” you can use function “graycoprops” that will calculate the statistics properties defined in its “properties” argument. For more information refer to below link:

https://www.mathworks.com/help/images/ref/graycoprops.html

Since you have already extracted features. For classification using SVM using custom kernel please refer to below link:

https://www.mathworks.com/help/stats/support-vector-machines-for-binary-classification.html