Python Image Processing Cookbook

How to do it...

Execute the following steps to implement the CLS filter and to compare restored images with different frequency domain filters:

1. First, define the cls_filter() function to implement the constrained least squares filter:

def cls_filter(y,h,c,lambd):    Hf = fp.fft2(fp.ifftshift(h))    Cf = fp.fft2(fp.ifftshift(c))    Hf = np.conj(Hf) / (Hf*np.conj(Hf) + lambd*Cf*np.conj(Cf))    Yf = fp.fft2(y)    I = Yf*Hf     im = np.abs(fp.ifft2(I))    return (im, Hf) 

2. Read the input image and convert it into grayscale. Blur the input grayscale image using box-kernel (done in the frequency domain) and degrade the image using White Gaussian Noise (WGN) with a given standard deviation (σ):

x = rgb2gray(imread('images/building.png'))M, N = x.shapeh = np.ones((4,4))/16 ...