scikit image - How do I mention the direction of neighbours to calculate the glcm in skimage/Python? -

i'd trouble understanding angle parameter of greycomatrix in skimage. in example mentioned in documentation compute glcm's pixel right , up, mention 4 angles. , 4 glcm's.

>>> image = np.array([[0, 0, 1, 1], ...                   [0, 0, 1, 1], ...                   [0, 2, 2, 2], ...                   [2, 2, 3, 3]], dtype=np.uint8) >>> result = greycomatrix(image, [1], [0, np.pi/4, np.pi/2, 3*np.pi/4], levels=4) 

what should parameters pixel right , down?

there's typo in example included in documentation of greycomatrix (emphasis mine):

examples

compute 2 glcms: 1 1-pixel offset right, , 1 1-pixel offset upwards.

>>> image = np.array([[0, 0, 1, 1],   ...                   [0, 0, 1, 1], ...                   [0, 2, 2, 2], ...                   [2, 2, 3, 3]], dtype=np.uint8) >>> result = greycomatrix(image, [1], [0, np.pi/4, np.pi/2, 3*np.pi/4], ...                       levels=4) 

indeed, result contains four different glcm's rather two. these 4 matrices correspond possible combinations of 1 distance , 4 angles. calculate glcm corresponding "1-pixel offset right" distance , angle values should 1 , 0, respectively:

result = greycomatrix(image, distances=[1], angles=[0], levels=4) 

whereas calculate glcm corresponding "1-pixel offset upwards" parameters should 1 , np.pi/2:

result = greycomatrix(image, distances=[1], angles=[np.pi/2], levels=4) 

in example, distances=[1] , angles=[0, np.pi/4, np.pi/2, 3*np.pi/4]. select particular glcm 1 has specify appropriate indices angles , distances. thus, 1-pixel right glcm result[:, :, 0, 0] , 1-pixel upwards glcm result[:, :, 0, 2].

finally, if wish calculate "1-pixel offset downwards" glcm (↓) have transpose "1-pixel offset upwards" glcm (↑). important note in cases both glcm's similar. in fact, can ignore order of co-occurring intensities setting parameter symmetric true in call greycomatrix. doing so, al glcm's returned greycomatrix symmetric.