Test Binarizer

In [29]:

    

import numpy as np
import cv2
import sys
import os
sys.path.insert(0, os.path.abspath('..'))

import salientregions as sr
import scipy.io as sio


    

In [30]:

    

%pylab inline


    

    




Populating the interactive namespace from numpy and matplotlib

In [31]:

    

#Load the image
path_to_image = '../tests/images/Gray/Gray_scale.png'
img = cv2.imread(path_to_image, cv2.IMREAD_GRAYSCALE)
sr.show_image(img)


    

In [32]:

    

#Check if the image is actually grayscale
path_to_image = '../tests/images/Gray/Gray_scale.png'
img2 = cv2.imread(path_to_image)
print(sr.image_diff(img2[:,:,0], img, visualize=False))
print(sr.image_diff(img2[:,:,1], img, visualize=False))
print(sr.image_diff(img2[:,:,2], img, visualize=False))


    

    




True
True
True

In [33]:

    

SE_size_factor = 0.02
lam_factor = 3
lam = lam_factor * int(np.floor(SE_size_factor * np.sqrt(img.size / np.pi)))
print(lam )
area_factor_large=0.001
area_factor_verylarge=0.01
weights=(0.33, 0.33, 0.33)
offset=80
stepsize=1
connectivity=8


    

    




24

In [34]:

    

binarizer = sr.DatadrivenBinarizer(lam=lam, area_factor_large=area_factor_large, 
                                   area_factor_verylarge=area_factor_verylarge,
                                  weights=weights,
                                  offset=offset,
                                  stepsize=stepsize,
                                  connectivity=connectivity)


    

In [35]:

    

pylab.rcParams['figure.figsize'] = (10., 5.)
t, binarized, scores = binarizer.binarize_withthreshold(img, output_scores=True, visualize=True)


    

In [36]:

    

print("threshold: ", t, "true threshold:", 142)


    

    




('threshold: ', 142, 'true threshold:', 142)

In [37]:

    

binarized_true = cv2.imread('../tests/images/Gray/Binarization_data_driven.png', cv2.IMREAD_GRAYSCALE)
sr.show_image(binarized_true)
sr.image_diff(binarized, binarized_true)


    

    













    













    
Out[37]:





True