RandomSolarize in PyTorch

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

*Memos:

My post explains RandomInvert().
My post explains OxfordIIITPet().

RandomSolarize() can randomly solarize an image with a given probability as shown below:

*Memos:

The 1st argument for initialization is threshold(Required-Type:int or float). *All pixels equal or above this value are inverted.
The 2nd argument for initialization is p(Optional-Default:0.5-Type:int or float): *Memos:
- It's the probability of whether an image is solarized 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 RandomSolarize

rs = RandomSolarize(threshold=0)
rs = RandomSolarize(threshold=0, p=0.5)

rs
# RandomSolarize(p=0.5, threshold=0)

rs.threshold
# 0

rs.p
# 0.5

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

t300p1_data = OxfordIIITPet( # `t` is threshold.
    root="data",
    transform=RandomSolarize(threshold=300, p=1)
)

t256p1_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=256, p=1)
)

t250p1_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=250, p=1)
)

t240p1_data  = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=240, p=1)
)

t220p1_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=220, p=1)
)

t200p1_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=200, p=1)
)

t150p1_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=150, p=1)
)

t100p1_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=100, p=1)
)

t50p1_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=50, p=1)
)

t10p1_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=10, p=1)
)

t0p1_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=0, p=1)
)

tn10p1_data = OxfordIIITPet( # `n` is negative.
    root="data",
    transform=RandomSolarize(threshold=-10, p=1)
)

tn100p1_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=-100, p=1)
)

t0p0_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=0, p=0)
)

t0p05_data = OxfordIIITPet(
    root="data",
    transform=RandomSolarize(threshold=0, p=0.5)
    # transform=RandomSolarize(threshold=0)
)

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=t300p1_data, main_title="t300p1_data")
show_images1(data=t256p1_data, main_title="t256p1_data")
show_images1(data=t255p1_data, main_title="t255p1_data")
show_images1(data=t250p1_data, main_title="t250p1_data")
show_images1(data=t240p1_data, main_title="t240p1_data")
show_images1(data=t220p1_data, main_title="t220p1_data")
show_images1(data=t200p1_data, main_title="t200p1_data")
show_images1(data=t150p1_data, main_title="t150p1_data")
show_images1(data=t100p1_data, main_title="t100p1_data")
show_images1(data=t50p1_data, main_title="t50p1_data")
show_images1(data=t10p1_data, main_title="t10p1_data")
show_images1(data=t0p1_data, main_title="t0p1_data")
show_images1(data=tn10p1_data, main_title="tn10p1_data")
show_images1(data=tn100p1_data, main_title="tn100p1_data")
print()
show_images1(data=t0p0_data, main_title="t0p0_data")
show_images1(data=t0p0_data, main_title="t0p0_data")
show_images1(data=t0p0_data, main_title="t0p0_data")
print()
show_images1(data=t0p05_data, main_title="t0p05_data")
show_images1(data=t0p05_data, main_title="t0p05_data")
show_images1(data=t0p05_data, main_title="t0p05_data")
print()
show_images1(data=t0p1_data, main_title="t0p1_data")
show_images1(data=t0p1_data, main_title="t0p1_data")
show_images1(data=t0p1_data, main_title="t0p1_data")

# ↓ ↓ ↓ ↓ ↓ ↓ The code below is identical to the code above. ↓ ↓ ↓ ↓ ↓ ↓
def show_images2(data, main_title=None, t=None, prob=0):
    plt.figure(figsize=[10, 5])
    plt.suptitle(t=main_title, y=0.8, fontsize=14)
    if main_title != "origin_data":
        for i, (im, _) in zip(range(1, 6), data):
            plt.subplot(1, 5, i)
            rs = RandomSolarize(threshold=t, p=prob)
            plt.imshow(X=rs(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="t300p1_data", t=300, prob=1)
show_images2(data=origin_data, main_title="t256p1_data", t=256, prob=1)
show_images2(data=origin_data, main_title="t255p1_data", t=255, prob=1)
show_images2(data=origin_data, main_title="t250p1_data", t=250, prob=1)
show_images2(data=origin_data, main_title="t240p1_data", t=240, prob=1)
show_images2(data=origin_data, main_title="t220p1_data", t=220, prob=1)
show_images2(data=origin_data, main_title="t200p1_data", t=200, prob=1)
show_images2(data=origin_data, main_title="t150p1_data", t=150, prob=1)
show_images2(data=origin_data, main_title="t100p1_data", t=100, prob=1)
show_images2(data=origin_data, main_title="t50p1_data", t=50, prob=1)
show_images2(data=origin_data, main_title="t10p1_data", t=10, prob=1)
show_images2(data=origin_data, main_title="t0p1_data", t=0, prob=1)
show_images2(data=origin_data, main_title="tn10p1_data", t=-10, prob=1)
show_images2(data=origin_data, main_title="tn100p1_data", t=-100, prob=1)
print()
show_images2(data=origin_data, main_title="t0p0_data", t=0, prob=0)
show_images2(data=origin_data, main_title="t0p0_data", t=0, prob=0)
show_images2(data=origin_data, main_title="t0p0_data", t=0, prob=0)
print()
show_images2(data=origin_data, main_title="t0p05_data", t=0, prob=0.5)
show_images2(data=origin_data, main_title="t0p05_data", t=0, prob=0.5)
show_images2(data=origin_data, main_title="t0p05_data", t=0, prob=0.5)
print()
show_images2(data=origin_data, main_title="t0p1_data", t=0, prob=1)
show_images2(data=origin_data, main_title="t0p1_data", t=0, prob=1)
show_images2(data=origin_data, main_title="t0p1_data", t=0, 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 (2025-03-11T22:04:46+00:00) RandomSolarize in PyTorch. Retrieved from https://www.scien.cx/2025/03/11/randomsolarize-in-pytorch/

MLA

" » RandomSolarize in PyTorch." Super Kai (Kazuya Ito) | Sciencx - Tuesday March 11, 2025, https://www.scien.cx/2025/03/11/randomsolarize-in-pytorch/

HARVARD

Super Kai (Kazuya Ito) | Sciencx Tuesday March 11, 2025 » RandomSolarize in PyTorch., viewed ,<https://www.scien.cx/2025/03/11/randomsolarize-in-pytorch/>

VANCOUVER

Super Kai (Kazuya Ito) | Sciencx - » RandomSolarize in PyTorch. [Internet]. [Accessed ]. Available from: https://www.scien.cx/2025/03/11/randomsolarize-in-pytorch/

CHICAGO

" » RandomSolarize in PyTorch." Super Kai (Kazuya Ito) | Sciencx - Accessed . https://www.scien.cx/2025/03/11/randomsolarize-in-pytorch/

IEEE

" » RandomSolarize in PyTorch." Super Kai (Kazuya Ito) | Sciencx [Online]. Available: https://www.scien.cx/2025/03/11/randomsolarize-in-pytorch/. [Accessed: ]

rf:citation

» RandomSolarize in PyTorch | Super Kai (Kazuya Ito) | Sciencx | https://www.scien.cx/2025/03/11/randomsolarize-in-pytorch/ |

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