022-denoising.py

#!/usr/bin/env python
__author__ = "Sreenivas Bhattiprolu"
__license__ = "Feel free to copy, I appreciate if you acknowledge Python for Microscopists"

# https://www.youtube.com/watch?v=2PrzKWkqOtU


# Filters work by convolution with a moving window called a kernel.
#Convolution is nothing but multiplication of two arrays of different sizes. 
#The image will be of one size and the kernel with be of a different size, 
# #usually much smaller than image
# The input pixel is at the centre of the kernel. 
# The convolution is performed by sliding the kernel over the image, 
# $usually from top left of image.
# Linear filters and non-linear filters.
# Gaussian is an example of linear filter. 
#Non-linear filters preserve edges. 
#Median filter is an example of non-linear filter. 
#The algorithm selects the median value of all the pixels in the selected window
#NLM: https://scikit-image.org/docs/dev/auto_examples/filters/plot_nonlocal_means.html


### What is gaussian Kernel #####
import numpy 
from matplotlib import pyplot as plt

def gaussian_kernel(size, size_y=None):
    size = int(size)
    if not size_y:
        size_y = size
    else:
        size_y = int(size_y)
    x, y = numpy.mgrid[-size:size+1, -size_y:size_y+1]
    g = numpy.exp(-(x**2/float(size)+y**2/float(size_y)))
    return g / g.sum()
 

gaussian_kernel_array = gaussian_kernel(3)
print(gaussian_kernel_array)
plt.imshow(gaussian_kernel_array, cmap=plt.get_cmap('jet'), interpolation='nearest')
plt.colorbar()
plt.show()

############################ Denoising filters ###############
from skimage.restoration import denoise_nl_means, estimate_sigma
from skimage import img_as_ubyte, img_as_float
from matplotlib import pyplot as plt
from skimage import io
import numpy as np

img = img_as_float(io.imread("images/denoising/noisy_img.jpg"))
#Need to convert to float as we will be doing math on the array

from scipy import ndimage as nd
gaussian_img = nd.gaussian_filter(img, sigma=3)
plt.imsave("images/gaussian.jpg", gaussian_img)


median_img = nd.median_filter(img, size=3)
plt.imsave("images/median.jpg", median_img)

gaussian_img = nd.gaussian_filter(img, sigma=3)
plt.imsave("images/gaussian.jpg", gaussian_img)


##### NLM#####

sigma_est = np.mean(estimate_sigma(img, multichannel=True))

patch_kw = dict(patch_size=5,      
                patch_distance=3,  
                multichannel=True)

denoise_img = denoise_nl_means(img, h=1.15 * sigma_est, fast_mode=False,
                               patch_size=5, patch_distance=3, multichannel=True)
"""
denoise_img = denoise_nl_means(img, h=1.15 * sigma_est, fast_mode=False,
                           **patch_kw)
"""
denoise_img_as_8byte = img_as_ubyte(denoise_img)

plt.imshow(denoise_img)
#plt.imshow(denoise_img_as_8byte, cmap=plt.cm.gray, interpolation='nearest')
plt.imsave("images/NLM.jpg",denoise_img)