I have 1000 images and the pixel data for each of these images (64x128). I want to create a krigging model with the x values as the image data and the y values as the images' corresponding compliance values (which I have). So, I want to train with the image data & compliance values and ultimatley have the krigging model predict the compliance values for new images
But, with krigging, usually you have variables like x1, x2, etc. and then one y variable. What would my x variables be in this case? Would it be x1= the first pixel of every image? and x2 the 2nd and so on so that I have 64x128 variables? or would x1 be the 64x128 matrix of image 1, and x2 of image 2, and so on. not sure. thanks!