RandomPosterize in PyTorch

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

*Memos:

My post explains OxfordIIITPet().

RandomPosterize() can randomly posterize an image with a given probability as shown below:

*Memos:

The 1st argument for initialization is bits(Required-Type:int): *Memos:
- It's the number of bits to keep for each channel.
- It must be x <= 8.
The 1st argument for initialization is p(Optional-Default:0.5-Type:int or float): *Memos:
- It's the probability of whether an image is posterized or not.
- It must be 0 <= x <= 1.
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 RandomPosterize

randomposterize = RandomPosterize(bits=1)
randomposterize = RandomPosterize(bits=1, p=0.5)

randomposterize 
# RandomPosterize(p=0.5, bits=1)

randomposterize.bits
# 1

randomposterize.p
# 0.5

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

b8p1origin_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=8, p=1)
)

b7p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=7, p=1)
)

b6p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=6, p=1)
)

b5p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=5, p=1)
)

b4p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=4, p=1)
)

b3p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=3, p=1)
)

b2p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=2, p=1)
)

b1p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=1, p=1)
)

b0p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=0, p=1)
)

bn1p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=-1, p=1)
)

bn10p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=-10, p=1)
)

bn100p1_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=-100, p=1)
)

b1p0_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=1, p=0)
)

b1p05_data = OxfordIIITPet(
    root="data",
    transform=RandomPosterize(bits=1, p=0.5)
    # transform=RandomPosterize(bits=1)
)

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")
print()
show_images1(data=b8p1origin_data, main_title="b8p1origin_data")
show_images1(data=b7p1_data, main_title="b7p1_data")
show_images1(data=b6p1_data, main_title="b6p1_data")
show_images1(data=b5p1_data, main_title="b5p1_data")
show_images1(data=b4p1_data, main_title="b4p1_data")
show_images1(data=b3p1_data, main_title="b3p1_data")
show_images1(data=b2p1_data, main_title="b2p1_data")
show_images1(data=b1p1_data, main_title="b1p1_data")
show_images1(data=b0p1_data, main_title="b0p1_data")
show_images1(data=bn1p1_data, main_title="bn1p1_data")
show_images1(data=bn10p1_data, main_title="bn10p1_data")
show_images1(data=bn100p1_data, main_title="bn100p1_data")
print()
show_images1(data=b1p0_data, main_title="b1p0_data")
show_images1(data=b1p0_data, main_title="b1p0_data")
show_images1(data=b1p0_data, main_title="b1p0_data")
print()
show_images1(data=b1p05_data, main_title="b1p05_data")
show_images1(data=b1p05_data, main_title="b1p05_data")
show_images1(data=b1p05_data, main_title="b1p05_data")
print()
show_images1(data=b1p1_data, main_title="b1p1_data")
show_images1(data=b1p1_data, main_title="b1p1_data")
show_images1(data=b1p1_data, main_title="b1p1_data")

# ↓ ↓ ↓ ↓ ↓ ↓ The code below is identical to the code above. ↓ ↓ ↓ ↓ ↓ ↓
def show_images2(data, main_title=None, b=None, prob=0):
    plt.figure(figsize=[10, 5])
    plt.suptitle(t=main_title, y=0.8, fontsize=14)
    if b != None:
        for i, (im, _) in zip(range(1, 6), data):
            plt.subplot(1, 5, i)
            rp = RandomPosterize(bits=b, p=prob)
            plt.imshow(X=rp(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")
print()
show_images2(data=origin_data, main_title="b8p1origin_data", b=8, prob=1)
show_images2(data=origin_data, main_title="b7p1_data", b=7, prob=1)
show_images2(data=origin_data, main_title="b6p1_data", b=6, prob=1)
show_images2(data=origin_data, main_title="b5p1_data", b=5, prob=1)
show_images2(data=origin_data, main_title="b4p1_data", b=4, prob=1)
show_images2(data=origin_data, main_title="b3p1_data", b=3, prob=1)
show_images2(data=origin_data, main_title="b2p1_data", b=2, prob=1)
show_images2(data=origin_data, main_title="b1p1_data", b=1, prob=1)
show_images2(data=origin_data, main_title="b0p1_data", b=0, prob=1)
show_images2(data=origin_data, main_title="bn1p1_data", b=-1, prob=1)
show_images2(data=origin_data, main_title="bn10p1_data", b=-10, prob=1)
show_images2(data=origin_data, main_title="bn100p1_data", b=-100, prob=1)
print()
show_images2(data=origin_data, main_title="b1p0_data", b=1, prob=0)
show_images2(data=origin_data, main_title="b1p0_data", b=1, prob=0)
show_images2(data=origin_data, main_title="b1p0_data", b=1, prob=0)
print()
show_images2(data=origin_data, main_title="b1p05_data", b=1, prob=0.5)
show_images2(data=origin_data, main_title="b1p05_data", b=1, prob=0.5)
show_images2(data=origin_data, main_title="b1p05_data", b=1, prob=0.5)
print()
show_images2(data=origin_data, main_title="b1p1_data", b=1, prob=1)
show_images2(data=origin_data, main_title="b1p1_data", b=1, prob=1)
show_images2(data=origin_data, main_title="b1p1_data", b=1, prob=1)

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 (2025-02-17T01:59:34+00:00) RandomPosterize in PyTorch. Retrieved from https://www.scien.cx/2025/02/17/randomposterize-in-pytorch-2/

MLA

" » RandomPosterize in PyTorch." Super Kai (Kazuya Ito) | Sciencx - Monday February 17, 2025, https://www.scien.cx/2025/02/17/randomposterize-in-pytorch-2/

HARVARD

Super Kai (Kazuya Ito) | Sciencx Monday February 17, 2025 » RandomPosterize in PyTorch., viewed ,<https://www.scien.cx/2025/02/17/randomposterize-in-pytorch-2/>

VANCOUVER

Super Kai (Kazuya Ito) | Sciencx - » RandomPosterize in PyTorch. [Internet]. [Accessed ]. Available from: https://www.scien.cx/2025/02/17/randomposterize-in-pytorch-2/

CHICAGO

" » RandomPosterize in PyTorch." Super Kai (Kazuya Ito) | Sciencx - Accessed . https://www.scien.cx/2025/02/17/randomposterize-in-pytorch-2/

IEEE

" » RandomPosterize in PyTorch." Super Kai (Kazuya Ito) | Sciencx [Online]. Available: https://www.scien.cx/2025/02/17/randomposterize-in-pytorch-2/. [Accessed: ]

rf:citation

» RandomPosterize in PyTorch | Super Kai (Kazuya Ito) | Sciencx | https://www.scien.cx/2025/02/17/randomposterize-in-pytorch-2/ |

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