there are two pipeline functions separately for SIFT and HOG feature descriptors to match a given input image with a specified input hyperparameter to the original text.
The input cropped images may vary in brightness, rotation, scale, and resolution.