notebook.community

Edit and run



In [1]:

    
import cv2
import matplotlib.pyplot as plt
%matplotlib inline
import sys
import pdb
import numpy as np



In [2]:

    
from inca.models.image import Image



In [3]:

    
i = Image.get(849555)
pan = i.get_panchromatic_asset()
pan.download_asset()



In [4]:

    
# Get Pan Image
pan_im = cv2.imread(pan.local_asset_path,-1)
plt.imshow(pan_im[500:,0:500],cmap=plt.get_cmap('gray'))

# Get Predicted Heatmap
plt.figure()
probs,_,hm,cm =i.get_prediction_assets(10972)
hm.download_asset()
hm_im = cv2.imread(hm.local_asset_path)
plt.imshow(hm_im,cmap=plt.get_cmap('gray'))

#Get Ground Truth Mask (cant figure out how to get it, so getting from confusion matrix)
plt.figure()
cm.download_asset()
cm_im = cv2.imread(cm.local_asset_path)

#Get Raw probs
probs.download_asset()
print probs.local_asset_path









    



/tmp/d09f3a6a93ac4b4386b81042af43d46f.npz






    












    












    





<matplotlib.figure.Figure at 0x7f24566a4dd0>



In [5]:

    
test1_im = pan_im[500:1000,0:500]
test1_hm = hm_im[500:1000,0:500]
test1_cm = cm_im[500:1000,0:500]
probs_im = np.load(probs.local_asset_path)['arr_0']
test1_probs = probs_im[500:1000,0:500]



In [124]:

    
## CRF Realted stuff
import pydensecrf.densecrf as dcrf
from pydensecrf.utils import unary_from_labels, create_pairwise_bilateral, create_pairwise_gaussian, unary_from_softmax
d = dcrf.DenseCRF2D(500, 500, 1)
U = -np.log(test1_probs + 0.0000001)
U = U.reshape((1,-1))
print U.dtype
d.setUnaryEnergy(unary_from_softmax(U))
d.addPairwiseGaussian(sxy=(3,3), compat=3, kernel=dcrf.DIAG_KERNEL, normalization=dcrf.NORMALIZE_SYMMETRIC)


Q = d.inference(50)
map = np.argmax(Q, axis=0).reshape((500,500))
proba = np.array(Q)
print np.max(proba), np.min(proba)
plt.imshow(proba.reshape(500,500),cmap=plt.get_cmap('gray'))









    



float16






    



---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-124-e9ae7ec3fbfe> in <module>()
      8 d.setUnaryEnergy(unary_from_softmax(U))
      9 d.addPairwiseGaussian(sxy=(3,3), compat=3, kernel=dcrf.DIAG_KERNEL, normalization=dcrf.NORMALIZE_SYMMETRIC)
---> 10 feats = create_pairwise_bilateral(sdims=(80, 80), schan=(13, 13),img=test1_im, chdim=1)
     11 
     12 

/home/ubuntu/anaconda/lib/python2.7/site-packages/pydensecrf/utils.pyc in create_pairwise_bilateral(sdims, schan, img, chdim)
    157     # scale mesh accordingly
    158     for i, s in enumerate(sdims):
--> 159         mesh[i] /= s
    160 
    161     feats = np.concatenate([mesh, im_feat])

IndexError: index 1 is out of bounds for axis 0 with size 1



In [8]:

    
import pydensecrf.densecrf as dcrf
from pydensecrf.utils import unary_from_labels, create_pairwise_bilateral, create_pairwise_gaussian, unary_from_softmax, softmax_to_unary



In [7]:

    
test1_probs.shape









    Out[7]:





(500, 500)



In [9]:

    
unary = softmax_to_unary(test1_probs)









    



WARNING:root:pydensecrf.softmax_to_unary is deprecated, use unary_from_softmax instead.
/home/ubuntu/anaconda/lib/python2.7/site-packages/pydensecrf/utils.py:79: RuntimeWarning: divide by zero encountered in log
  return -np.log(sm).reshape([num_cls, -1]).astype(np.float32)



In [10]:

    
unary = np.ascontiguousarray(unary)



In [13]:

    
d = dcrf.DenseCRF(500,500, 1)



In [14]:

    
d.setUnaryEnergy(unary)



In [25]:

    
feats = create_pairwise_gaussian(sdims=(3, 3),shape=test1_im.shape[1:])









    



---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-25-c4cd1a875dbc> in <module>()
----> 1 feats = create_pairwise_gaussian(sdims=(3, 3),shape=test1_im.shape[1:])

/home/ubuntu/anaconda/lib/python2.7/site-packages/pydensecrf/utils.pyc in create_pairwise_gaussian(sdims, shape)
    108     # scale mesh accordingly
    109     for i, s in enumerate(sdims):
--> 110         mesh[i] /= s
    111     return mesh.reshape([len(sdims), -1])
    112 

IndexError: index 1 is out of bounds for axis 0 with size 1



In [24]:

    
test1_im.shape[2:]









    Out[24]:





()



In [26]:

    
d.addPairwiseGaussian(sxy=(3,3), compat=3, kernel=dcrf.DIAG_KERNEL, normalization=dcrf.NORMALIZE_SYMMETRIC)









    



---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-26-25e8d6eb59d8> in <module>()
----> 1 d.addPairwiseGaussian(sxy=(3,3), compat=3, kernel=dcrf.DIAG_KERNEL, normalization=dcrf.NORMALIZE_SYMMETRIC)

AttributeError: 'pydensecrf.densecrf.DenseCRF' object has no attribute 'addPairwiseGaussian'



In [59]:

    
d2 = dcrf.DenseCRF2D(500, 500, 1)



In [60]:

    
unary = -np.log(test1_probs + 0.0000001)



In [61]:

    
unary = softmax_to_unary(test1_probs.reshape(1,-1))
unary = np.ascontiguousarray(unary)
d2.setUnaryEnergy(unary)









    



WARNING:root:pydensecrf.softmax_to_unary is deprecated, use unary_from_softmax instead.



In [62]:

    
d2.addPairwiseGaussian(sxy=3, compat=3)



In [71]:

    
Q=d.inference(50)



In [72]:

    
res = np.argmax(Q, axis=0).reshape((500,500))









    



---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-72-402ecae58b16> in <module>()
----> 1 res = np.argmax(Q, axis=0).reshape((500,500))

ValueError: cannot reshape array of size 500 into shape (500,500)



In [73]:

    
Q.shape









    



---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-73-afc0d1133397> in <module>()
----> 1 Q.shape

AttributeError: 'pydensecrf.eigen.MatrixXf' object has no attribute 'shape'



In [74]:

    
proba=np.array(Q)



In [75]:

    
proba.shape









    Out[75]:





(500, 500)



In [76]:

    
print np.max(proba)



In [80]:

    
proba_copy=np.copy(proba)



In [81]:

    
proba_copy=proba_copy/np.max(proba_copy)



In [82]:

    
np.max(proba_copy)









    Out[82]:





1.0



In [83]:

    
np.min(proba_copy)









    Out[83]:





0.0



In [84]:

    
proba_copy=proba_copy*255.0



In [86]:

    
plt.imshow(proba_copy)









    Out[86]:





<matplotlib.image.AxesImage at 0x7f24507a4310>



In [58]:

    
plt.imshow(test1_im,cmap=plt.get_cmap('gray'))









    Out[58]:





<matplotlib.image.AxesImage at 0x7f2456944b90>



In [53]:

    
d3 = dcrf.DenseCRF(500*500, 1)



In [54]:

    
unary = softmax_to_unary(test1_probs.reshape(1,-1))
unary = np.ascontiguousarray(unary)
d3.setUnaryEnergy(unary)









    



WARNING:root:pydensecrf.softmax_to_unary is deprecated, use unary_from_softmax instead.



In [55]:

    
d3.addPairwiseGaussian(sxy=3, compat=3)









    



---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-55-de0914638f95> in <module>()
----> 1 d3.addPairwiseGaussian(sxy=3, compat=3)

AttributeError: 'pydensecrf.densecrf.DenseCRF' object has no attribute 'addPairwiseGaussian'



In [87]:

    
from pydensecrf.utils import unary_from_softmax, create_pairwise_bilateral



In [91]:

    
pairwise_energy = create_pairwise_bilateral(sdims=(80,80), schan=(13,), img=test1_im, chdim=-1)



In [88]:

    
create_pairwise_bilateral?



In [89]:

    
test1_im.shape









    Out[89]:





(500, 500)



In [93]:

    
d2.addPairwiseEnergy(pairwise_energy,compat=10)



In [94]:

    
Q=d.inference(50)



In [95]:

    
proba=np.array(Q)



In [96]:

    
plt.imshow(proba)









    Out[96]:





<matplotlib.image.AxesImage at 0x7f245070d910>



In [97]:

    
np.max(proba)









    Out[97]:





0.19316731



In [98]:

    
aa=test1_probs.reshape((1,-1))



In [99]:

    
aa.shape









    Out[99]:





(1, 250000)



In [ ]: