[FIXED] How to convert image in fast fourier transform signal?

Issue

I am trying to convert image into fast fourier transform signal and used the following peace of code:

fake_A1 = tf.signal.fft2d(fake_A1)

where input image type is: <class 'numpy.ndarray'> but I am getting following error:

tensorflow.python.framework.errors_impl.InvalidArgumentError: Value for attr 'Tcomplex' of float is not in the list of allowed values: complex64, complex128
; NodeDef: {{node FFT2D}}; Op<name=FFT2D; signature=input:Tcomplex -> output:Tcomplex; attr=Tcomplex:type,default=DT_COMPLEX64,allowed=[DT_COMPLEX64, DT_COMPLEX128]> [Op:FFT2D]

How may I make it solve?

Solution

P.S.: If you want to make edits then do it on your question, not as an answer.

Now coming to the topic: 2D FFT of an image. Firstly an image is of shape:

image.shape = (3,rows,columns)

Where 3 stands for 3 matrices which are of 2 dimensions, corresponding to RGB. Hence to carry out 2D FFT, we first need to flatten this by converting it to grayscale. I found a useful tutorial here on ThePythonCodingBook. Ill add the code here for TL:DR purposes.

import matplotlib.pyplot as plt
image_filename = "Earth.png"
# Read and process image
image = plt.imread(image_filename)
image = image[:, :, :3].mean(axis=2)  # Convert to grayscale
print(image.shape)
plt.set_cmap("gray")
plt.imshow(image)
plt.axis("off")
plt.show()

import numpy as np
import matplotlib.pyplot as plt
image_filename = "Earth.png"
def calculate_2dft(input):
    ft = np.fft.ifftshift(input)
    ft = np.fft.fft2(ft)
    return np.fft.fftshift(ft)
# Read and process image
image = plt.imread(image_filename)
image = image[:, :, :3].mean(axis=2)  # Convert to grayscale
plt.set_cmap("gray")
ft = calculate_2dft(image)
plt.subplot(121)
plt.imshow(image)
plt.title('Grayscale Image')
plt.axis("off")
plt.subplot(122)
plt.imshow(np.log(abs(ft)))
plt.title("2D FFT")
plt.axis("off")
plt.show()

I'll add the grayscale and FFT plots as well.

Hope this helps.

Answered By - Gautam Chettiar