Blur image using GaussianBlur operator

import torch
import kornia
import cv2
import numpy as np

import matplotlib.pyplot as plt

# read the image with OpenCV
img: np.ndarray = cv2.imread('./data/lena.jpg')
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

# convert to torch tensor
data: torch.tensor = kornia.image_to_tensor(img, keepdim=False)  # BxCxHxW

# create the operator
gauss = kornia.filters.GaussianBlur2d((11, 11), (10.5, 10.5))

# blur the image
x_blur: torch.tensor = gauss(data.float())

# convert back to numpy
img_blur: np.ndarray = kornia.tensor_to_image(x_blur.byte())

# Create the plot
fig, axs = plt.subplots(1, 2, figsize=(16, 10))
axs = axs.ravel()

axs[0].axis('off')
axs[0].set_title('image source')
axs[0].imshow(img)

axs[1].axis('off')
axs[1].set_title('image blurred')
axs[1].imshow(img_blur)