In [1]:

    

%matplotlib inline
%pylab inline


    

    




Populating the interactive namespace from numpy and matplotlib

In [10]:

    

import menpo.io as mio
from menpo.landmark import labeller, ibug_face_66
from menpofast.utils import convert_from_menpo

path = '/data/'
group = 'ibug_face_66'

training_images_color = []
training_images = []
for i in mio.import_images(path + 'PhD/DataBases/faces/ibug/', verbose=True, 
                           max_images=None):
    
    # convert the image from menpo Image to menpofast Image (channels at front)
    i = convert_from_menpo(i)
    
    labeller(i, 'PTS', eval(group))
    i.crop_to_landmarks_proportion_inplace(0.5, group='PTS')
    i = i.rescale_landmarks_to_diagonal_range(200, group=group)
    
    training_images_color.append(i)
    
    if i.n_channels == 3:
        ii = i.as_greyscale(mode='average')
    training_images.append(ii)


    

    




- Loading 135 assets: [====================] 100%

In [11]:

    

from menpo.visualize import visualize_images

visualize_images(training_images_color, colours='r', linewidths=2)


    

In [12]:

    

img = training_images_color[98]

img.view_widget(colours='r', linewidths=2)


    

In [16]:

    

from menpofast.feature import fast_dsift

img = training_images[98]

masked_img = img.as_masked()
masked_img.build_mask_around_landmarks((34, 34), group=group)

dsift_img = fast_dsift(masked_img)

dsift_img.view_widget(colours='r', linewidths=2)


    

In [ ]:

    

import menpo.io as mio
from menpo.landmark import labeller, streetscene_car_view_1
from menpofast.utils import convert_from_menpo

training_images_color = []
training_images = []
for i in mio.import_images('/data/PhD/DataBases/cars/cmu_car_data1/view1/', verbose=True, 
                           max_images=20):
    
    # convert the image from menpo Image to menpofast Image (channels at front)
    i = convert_from_menpo(i)
    
    labeller(i, 'PTS', streetscene_car_view_1)
    i.crop_to_landmarks_proportion_inplace(0.5, group='streetscene_car_view_1')
    i = i.rescale_landmarks_to_diagonal_range(200, group='streetscene_car_view_1')
    
    training_images_color.append(i)
    
    if i.n_channels == 3:
        ii = i.as_greyscale(mode='average')
    training_images.append(ii)


    

In [ ]:

    

from menpo.visualize import visualize_images

visualize_images(training_images_color, colours='r', linewidths=2)


    

In [ ]:

    

img = training_images_color[11]

img.view_widget(colours='r', linewidths=2)


    

In [ ]:

    

from menpofast.feature import aam_dsift

img = training_images[11]

masked_img = img.as_masked()
masked_img.build_mask_around_landmarks((34, 34), group='streetscene_car_view_1')

dsift_img = aam_dsift(masked_img)

dsift_img.view_widget(colours='r', linewidths=2)


    

In [3]:

    

from scipy.signal import cosine

height = 100
width = 100

cosine_mask = np.sum(np.meshgrid(cosine(height), cosine(width)), axis=0)


    

In [4]:

    

imshow(cosine_mask)


    

    
Out[4]:





<matplotlib.image.AxesImage at 0x7fecc758a850>






    

In [21]:

    

from __future__ import division

step = 2

w1 = np.cos(np.linspace(-pi/step, pi/step, height))
w2 = np.cos(np.linspace(-pi/step, pi/step, width))

w = w1[..., None].dot(w2[None, ...])


    

In [22]:

    

imshow(w)


    

    
Out[22]:





<matplotlib.image.AxesImage at 0x7fecc6ebac10>






    

In [20]:

    

w1


    

    
Out[20]:





array([  6.12323400e-17,   3.17279335e-02,   6.34239197e-02,
         9.50560433e-02,   1.26592454e-01,   1.58001396e-01,
         1.89251244e-01,   2.20310533e-01,   2.51147987e-01,
         2.81732557e-01,   3.12033446e-01,   3.42020143e-01,
         3.71662456e-01,   4.00930535e-01,   4.29794912e-01,
         4.58226522e-01,   4.86196736e-01,   5.13677392e-01,
         5.40640817e-01,   5.67059864e-01,   5.92907929e-01,
         6.18158986e-01,   6.42787610e-01,   6.66769001e-01,
         6.90079011e-01,   7.12694171e-01,   7.34591709e-01,
         7.55749574e-01,   7.76146464e-01,   7.95761841e-01,
         8.14575952e-01,   8.32569855e-01,   8.49725430e-01,
         8.66025404e-01,   8.81453363e-01,   8.95993774e-01,
         9.09631995e-01,   9.22354294e-01,   9.34147860e-01,
         9.45000819e-01,   9.54902241e-01,   9.63842159e-01,
         9.71811568e-01,   9.78802446e-01,   9.84807753e-01,
         9.89821442e-01,   9.93838464e-01,   9.96854776e-01,
         9.98867339e-01,   9.99874128e-01,   9.99874128e-01,
         9.98867339e-01,   9.96854776e-01,   9.93838464e-01,
         9.89821442e-01,   9.84807753e-01,   9.78802446e-01,
         9.71811568e-01,   9.63842159e-01,   9.54902241e-01,
         9.45000819e-01,   9.34147860e-01,   9.22354294e-01,
         9.09631995e-01,   8.95993774e-01,   8.81453363e-01,
         8.66025404e-01,   8.49725430e-01,   8.32569855e-01,
         8.14575952e-01,   7.95761841e-01,   7.76146464e-01,
         7.55749574e-01,   7.34591709e-01,   7.12694171e-01,
         6.90079011e-01,   6.66769001e-01,   6.42787610e-01,
         6.18158986e-01,   5.92907929e-01,   5.67059864e-01,
         5.40640817e-01,   5.13677392e-01,   4.86196736e-01,
         4.58226522e-01,   4.29794912e-01,   4.00930535e-01,
         3.71662456e-01,   3.42020143e-01,   3.12033446e-01,
         2.81732557e-01,   2.51147987e-01,   2.20310533e-01,
         1.89251244e-01,   1.58001396e-01,   1.26592454e-01,
         9.50560433e-02,   6.34239197e-02,   3.17279335e-02,
         6.12323400e-17])

In [23]:

    

np.pi


    

    
Out[23]:





3.141592653589793

In [ ]: