GaussianBlur in PyTorch (1)

This content originally appeared on DEV Community and was authored by Super Kai (Kazuya Ito)

*Memos:

My post explains OxfordIIITPet().

GaussianBlur() can randomly blur an image as shown below:

*Memos:

The 1st argument for initialization is num_output_channels(Required-Type:int or tuple/list(int)): *Memos:
- It's [height, width].
- It must be odd 1 <= x.
- A tuple/list must be the 1D with 1 or 2 elements.
- A single value(int or tuple/list(int)) means [num_output_channels, num_output_channels].
The 2nd argument for initialization is sigma(Optional-Default:(0.1, 2.0)-Type:int or tuple/list(int)): *Memos:
- It's [min, max] so it must be min <= max.
- It must be 0 < x.
- A tuple/list must be the 1D with 1 or 2 elements.
- A single value(int or tuple/list(int)) means [sigma, sigma].
The 1st argument is img(Required-Type:PIL Image or tensor(int)): *Memos:
- A tensor must be 2D or 3D.
- Don't use img=.
v2 is recommended to use according to V1 or V2? Which one should I use?.

from torchvision.datasets import OxfordIIITPet
from torchvision.transforms.v2 import GaussianBlur

gaussianblur = GaussianBlur(kernel_size=1)
gaussianblur = GaussianBlur(kernel_size=1, sigma=(0.1, 2.0))

gaussianblur
# GaussianBlur(kernel_size=(1, 1), sigma=[0.1, 2.0])

gaussianblur.kernel_size 
# (1, 1)

gaussianblur.sigma
# [0.1, 2.0]

origin_data = OxfordIIITPet(
    root="data",
    transform=None
)

ks1_data = OxfordIIITPet( # `ks` is kernel_size.
    root="data",
    transform=GaussianBlur(kernel_size=1)
    # transform=GaussianBlur(kernel_size=[1])
    # transform=GaussianBlur(kernel_size=[1, 1])
)

ks5_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=5)
)

ks11_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=11)
)

ks51_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=51)
)

ks101_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=101)
)

ks5_51_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=[5, 51])
)

ks51_5_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=[51, 5])
)

ks1s01_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=1, sigma=0.1)
)

ks5s01_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=5, sigma=0.1)
)

ks11s01_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=11, sigma=0.1)
)

ks51s01_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=51, sigma=0.1)
)

ks101s01_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=101, sigma=0.1)
)

ks9_51s01_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=[9, 51], sigma=0.1)
)

ks51_9s01_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=[51, 9], sigma=0.1)
)

ks1s100_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=1, sigma=100)
)

ks5s100_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=5, sigma=100)
)

ks11s100_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=11, sigma=100)
)

ks51s100_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=51, sigma=100)
)

ks101s100_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=101, sigma=100)
)

ks9_51s100_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=[9, 51], sigma=100)
)

ks51_9s100_data = OxfordIIITPet(
    root="data",
    transform=GaussianBlur(kernel_size=[51, 9], sigma=100)
)

import matplotlib.pyplot as plt

def show_images1(data, main_title=None):
    plt.figure(figsize=[10, 5])
    plt.suptitle(t=main_title, y=0.8, fontsize=14)
    for i, (im, _) in zip(range(1, 6), data):
        plt.subplot(1, 5, i)
        plt.imshow(X=im)
        plt.xticks(ticks=[])
        plt.yticks(ticks=[])
    plt.tight_layout()
    plt.show()

show_images1(data=origin_data, main_title="origin_data")
show_images1(data=ks1_data, main_title="ks1_data")
show_images1(data=ks5_data, main_title="ks5_data")
show_images1(data=ks11_data, main_title="ks11_data")
show_images1(data=ks51_data, main_title="ks51_data")
show_images1(data=ks101_data, main_title="ks101_data")
show_images1(data=ks5_51_data, main_title="ks5_51_data")
show_images1(data=ks51_5_data, main_title="ks51_5_data")
print()
show_images1(data=origin_data, main_title="origin_data")
show_images1(data=ks1s01_data, main_title="ks1s01_data")
show_images1(data=ks5s01_data, main_title="ks5s01_data")
show_images1(data=ks11s01_data, main_title="ks11s01_data")
show_images1(data=ks51s01_data, main_title="ks51s01_data")
show_images1(data=ks101s01_data, main_title="ks101s01_data")
show_images1(data=ks9_51s01_data, main_title="ks9_51s01_data")
show_images1(data=ks51_9s01_data, main_title="ks51_9s01_data")
print()
show_images1(data=origin_data, main_title="origin_data")
show_images1(data=ks1s01_data, main_title="ks1s100_data")
show_images1(data=ks5s01_data, main_title="ks5s100_data")
show_images1(data=ks11s01_data, main_title="ks11s100_data")
show_images1(data=ks51s01_data, main_title="ks51s100_data")
show_images1(data=ks101s01_data, main_title="ks101s100_data")
show_images1(data=ks9_51s01_data, main_title="ks9_51s100_data")
show_images1(data=ks51_9s01_data, main_title="ks51_9s100_data")

# ↓ ↓ ↓ ↓ ↓ ↓ The code below is identical to the code above. ↓ ↓ ↓ ↓ ↓ ↓
def show_images2(data, main_title=None, ks=None, s=(0.1, 2.0)):
    plt.figure(figsize=[10, 5])
    plt.suptitle(t=main_title, y=0.8, fontsize=14)
    if ks:
        for i, (im, _) in zip(range(1, 6), data):
            plt.subplot(1, 5, i)
            gb = GaussianBlur(kernel_size=ks, sigma=s)
            plt.imshow(X=gb(im))
            plt.xticks(ticks=[])
            plt.yticks(ticks=[])
    else:
        for i, (im, _) in zip(range(1, 6), data):
            plt.subplot(1, 5, i)
            plt.imshow(X=im)
            plt.xticks(ticks=[])
            plt.yticks(ticks=[])
    plt.tight_layout()
    plt.show()

show_images2(data=origin_data, main_title="origin_data")
show_images2(data=origin_data, main_title="ks1_data", ks=1)
show_images2(data=origin_data, main_title="ks5_data", ks=5)
show_images2(data=origin_data, main_title="ks11_data", ks=11)
show_images2(data=origin_data, main_title="ks51_data", ks=51)
show_images2(data=origin_data, main_title="ks101_data", ks=101)
show_images2(data=origin_data, main_title="ks5_51data", ks=[5, 51])
show_images2(data=origin_data, main_title="ks51_5_data", ks=[51, 5])
print()
show_images2(data=origin_data, main_title="origin_data")
show_images2(data=origin_data, main_title="ks1s01_data", ks=1, s=0.1)
show_images2(data=origin_data, main_title="ks5s01_data", ks=5, s=0.1)
show_images2(data=origin_data, main_title="ks11s01_data", ks=11, s=0.1)
show_images2(data=origin_data, main_title="ks51s01_data", ks=51, s=0.1)
show_images2(data=origin_data, main_title="ks101s01_data", ks=101, s=0.1)
show_images2(data=origin_data, main_title="ks5_51s01data", ks=[5, 51])
show_images2(data=origin_data, main_title="ks51_5s01_data", ks=[51, 5])
print()
show_images2(data=origin_data, main_title="origin_data")
show_images2(data=origin_data, main_title="ks1s100_data", ks=1, s=100)
show_images2(data=origin_data, main_title="ks5s100_data", ks=5, s=100)
show_images2(data=origin_data, main_title="ks11s100_data", ks=11, s=100)
show_images2(data=origin_data, main_title="ks51s100_data", ks=51, s=100)
show_images2(data=origin_data, main_title="ks101s100_data", ks=101, s=100)
show_images2(data=origin_data, main_title="ks5_51s100data", ks=[5, 51],
             s=100)
show_images2(data=origin_data, main_title="ks51_5s100_data", ks=[51, 5],
             s=100)

This content originally appeared on DEV Community and was authored by Super Kai (Kazuya Ito)

Print Share Comment Cite Upload Translate Updates

APA

Super Kai (Kazuya Ito) | Sciencx (2025-02-15T22:38:22+00:00) GaussianBlur in PyTorch (1). Retrieved from https://www.scien.cx/2025/02/15/gaussianblur-in-pytorch-1/

MLA

" » GaussianBlur in PyTorch (1)." Super Kai (Kazuya Ito) | Sciencx - Saturday February 15, 2025, https://www.scien.cx/2025/02/15/gaussianblur-in-pytorch-1/

HARVARD

Super Kai (Kazuya Ito) | Sciencx Saturday February 15, 2025 » GaussianBlur in PyTorch (1)., viewed ,<https://www.scien.cx/2025/02/15/gaussianblur-in-pytorch-1/>

VANCOUVER

Super Kai (Kazuya Ito) | Sciencx - » GaussianBlur in PyTorch (1). [Internet]. [Accessed ]. Available from: https://www.scien.cx/2025/02/15/gaussianblur-in-pytorch-1/

CHICAGO

" » GaussianBlur in PyTorch (1)." Super Kai (Kazuya Ito) | Sciencx - Accessed . https://www.scien.cx/2025/02/15/gaussianblur-in-pytorch-1/

IEEE

" » GaussianBlur in PyTorch (1)." Super Kai (Kazuya Ito) | Sciencx [Online]. Available: https://www.scien.cx/2025/02/15/gaussianblur-in-pytorch-1/. [Accessed: ]

rf:citation

» GaussianBlur in PyTorch (1) | Super Kai (Kazuya Ito) | Sciencx | https://www.scien.cx/2025/02/15/gaussianblur-in-pytorch-1/ |

Please log in to upload a file.

There are no updates yet.
Click the Upload button above to add an update.

You must be logged in to translate posts. Please log in or register.

Related Posts