Explain the idea of image histogram. How is it used to guide threshold selection when binarizing an image? How histogram shape would indicate if thresholding is likely to be good or subdivision of images into sub-regions is needed (for different thresholds for each sub-image)