Function ihadamard

Synopse

Inverse Hadamard Transform.

F = iaihadamard(f)
- Output:
  - F: Image.
- Input:
  - f: Image.

Function code



In [1]:

    
import numpy as np

def ihadamard(f):
    import ia898.src as ia
    f = np.asarray(f).astype(np.float64)
    if len(f.shape) == 1: f = f[:, newaxis]
    (m, n) = f.shape
    A = ia.hadamardmatrix(m)
    if (n == 1):
        F = np.dot(np.transpose(A), f)
    else:
        B = ia.hadamardmatrix(n)
        F = np.dot(np.dot(np.transpose(A), f), B)
    return F

Examples



In [2]:

    
testing = (__name__ == "__main__")
if testing:
    ! jupyter nbconvert --to python ihadamard.ipynb
    import sys
    import os
    ia898path = os.path.abspath('../../')
    if ia898path not in sys.path:
        sys.path.append(ia898path)
    import ia898.src as ia
    
    import matplotlib.pyplot as plt
    import matplotlib.image as mpimg
    %matplotlib inline









    



[NbConvertApp] Converting notebook ihadamard.ipynb to python
[NbConvertApp] Writing 1619 bytes to ihadamard.py

Example 1



In [3]:

    
if testing:
    f = mpimg.imread('../data/cameraman.tif')
    F = ia.hadamard(f)
    g = ia.ihadamard(F)
    print(sum(sum(abs(f.astype(float)-g.astype(float)))))
    
    nb = ia.nbshow(3)
    nb.nbshow(f, 'Original')
    nb.nbshow(ia.normalize(F), 'hadamard of f')
    nb.nbshow(ia.normalize(g), 'ihadamard of F')
    nb.nbshow()









    



0.0






    




                                    
                    Original                     
                    hadamard of f                     
                    ihadamard of F

Measuring time:



In [4]:

    
if testing:
    f = mpimg.imread('../data/cameraman.tif')
    F = ia.hadamard(f)
    print('Computational time is:')
    %timeit ia.ihadamard(F)









    



Computational time is:
100 loops, best of 3: 7.63 ms per loop



In [ ]: