Using numpy to implement typical features of Digital Image Processing, including HOG, LBP, Haar.

from HOG import HOG

hog_descriptor = HOG(img,block_size=3)

hog_vector = hog_descriptor.hog_features()

print(hog_vector.shape)

from LBP import LBP

lbp = LBP(img)

features = lbp.extend_lbp(3,8) # features = lbp.original_lbp()

r_features = lbp.extend_lbp(3,8,rotation_sensitive=False)

vector = lbp.get_lbp_vector(features,8)

print(vector.shape,vector[:10])

from Haar import Haar

haar_features = Haar(img)

vector = haar_features.get_haar_features()

The Haar features may take more time to calculate.