RandomResizedCrop in PyTorch (2)

#python #pytorch #randomresizedcrop #v2

*Memos:

My post explains RandomResizedCrop() about size argument.
My post explains RandomResizedCrop() about ratio argument.
My post explains RandomResizedCrop() about size argument with scale=[0, 0] and ratio=[1, 1].
My post explains RandomResizedCrop() about scale argument with ratio=[1, 1].
My post explains RandomResizedCrop() about ratio argument with scale=[0, 0].
My post explains OxfordIIITPet().

RandomResizedCrop() can crop a random part of an image, then resize it to a given size as shown below:

from torchvision.datasets import OxfordIIITPet
from torchvision.transforms.v2 import RandomResizedCrop
from torchvision.transforms.functional import InterpolationMode

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

s1000sc0_0origin_data = OxfordIIITPet( # `s` is size and `sc` is scale.
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0, 0])
)

s1000sc0_1_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0, 1])
)

s1000sc0_05_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0, 0.5])
)

s1000sc05_1_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.5, 1])
)

s1000sc0001_0001_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.001, 0.001])
)

s1000sc001_001_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.01, 0.01])
)

s1000sc01_01_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.1, 0.1])
)

s1000sc02_02_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.2, 0.2])
)

s1000sc03_03_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.3, 0.3])
)

s1000sc04_04_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.4, 0.4])
)

s1000sc05_05_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.5, 0.5])
)

s1000sc06_06_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.6, 0.6])
)

s1000sc07_07_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.7, 0.7])
)

s1000sc08_08_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.8, 0.8])
)

s1000sc09_09_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[0.9, 0.9])
)

s1000sc1_1origin_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[1, 1])
)

s1000sc10_10origin_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[10, 10])
)

s1000sc100_100origin_data = OxfordIIITPet(
    root="data",
    transform=RandomResizedCrop(size=1000, scale=[100, 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.tight_layout()
    plt.show()

show_images1(data=origin_data, main_title="origin_data")
print()
show_images1(data=s1000sc0_0origin_data, main_title="s1000sc0_0origin_data")
show_images1(data=s1000sc0_1_data, main_title="s1000sc0_1_data")
show_images1(data=s1000sc0_05_data, main_title="s1000sc0_05_data")
show_images1(data=s1000sc05_1_data, main_title="s1000sc05_1_data")
print()
show_images1(data=s1000sc0_0origin_data, main_title="s1000sc0_0origin_data")
show_images1(data=s1000sc0001_0001_data, main_title="s1000sc0001_0001_data")
show_images1(data=s1000sc001_001_data, main_title="s1000sc001_001_data")
show_images1(data=s1000sc01_01_data, main_title="s1000sc01_01_data")
show_images1(data=s1000sc02_02_data, main_title="s1000sc02_02_data")
show_images1(data=s1000sc03_03_data, main_title="s1000sc03_03_data")
show_images1(data=s1000sc04_04_data, main_title="s1000sc04_04_data")
show_images1(data=s1000sc05_05_data, main_title="s1000sc05_05_data")
show_images1(data=s1000sc06_06_data, main_title="s1000sc06_06_data")
show_images1(data=s1000sc07_07_data, main_title="s1000sc07_07_data")
show_images1(data=s1000sc08_08_data, main_title="s1000sc08_08_data")
show_images1(data=s1000sc09_09_data, main_title="s1000sc09_09_data")
show_images1(data=s1000sc1_1origin_data, main_title="s1000sc1_1origin_data")
show_images1(data=s1000sc10_10origin_data,
             main_title="s1000sc10_10origin_data")
show_images1(data=s1000sc100_100origin_data,
             main_title="s1000sc100_100origin_data")

# ↓ ↓ ↓ ↓ ↓ ↓ The code below is identical to the code above. ↓ ↓ ↓ ↓ ↓ ↓ 
def show_images2(data, main_title=None, s=None, sc=(0.08, 1.0),
                 r=(0.75, 1.3333333333333333),
                 ip=InterpolationMode.BILINEAR, a=True):
    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)
        if s:
            rrc = RandomResizedCrop(size=s, scale=sc, # Here
                                    ratio=r, interpolation=ip,
                                    antialias=a)
            plt.imshow(X=rrc(im)) # Here
        else:
            plt.imshow(X=im)
    plt.tight_layout()
    plt.show()

show_images2(data=origin_data, main_title="origin_data")
print()
show_images2(data=origin_data, main_title="s1000sc0_0origin_data", s=1000,
             sc=[0, 0])
show_images2(data=origin_data, main_title="s1000sc0_1_data", s=1000, sc=[0, 1])
show_images2(data=origin_data, main_title="s1000sc0_05_data", s=1000,
             sc=[0, 0.5])
show_images2(data=origin_data, main_title="s1000sc05_1_data", s=1000,
             sc=[0.5, 1])
print()
show_images2(data=origin_data, main_title="s1000sc0_0origin_data", s=1000,
             sc=[0, 0])
show_images2(data=origin_data, main_title="s1000sc0001_0001_data", s=1000,
             sc=[0.001, 0.001])
show_images2(data=origin_data, main_title="s1000sc001_001_data", s=1000,
             sc=[0.01, 0.01])
show_images2(data=origin_data, main_title="s1000sc01_01_data", s=1000,
             sc=[0.1, 0.1])
show_images2(data=origin_data, main_title="s1000sc02_02_data", s=1000,
             sc=[0.2, 0.2])
show_images2(data=origin_data, main_title="s1000sc03_03_data", s=1000,
             sc=[0.3, 0.3])
show_images2(data=origin_data, main_title="s1000sc04_04_data", s=1000,
             sc=[0.4, 0.4])
show_images2(data=origin_data, main_title="s1000sc05_05_data", s=1000,
             sc=[0.5, 0.5])
show_images2(data=origin_data, main_title="s1000sc06_06_data", s=1000,
             sc=[0.6, 0.6])
show_images2(data=origin_data, main_title="s1000sc07_07_data", s=1000,
             sc=[0.7, 0.7])
show_images2(data=origin_data, main_title="s1000sc08_08_data", s=1000,
             sc=[0.8, 0.8])
show_images2(data=origin_data, main_title="s1000sc09_09_data", s=1000,
             sc=[0.9, 0.9])
show_images2(data=origin_data, main_title="s1000sc1_1origin_data", s=1000,
             sc=[1, 1])
show_images2(data=origin_data, main_title="s1000sc10_10origin_data", s=1000, 
             sc=[10, 10])
show_images2(data=origin_data, main_title="s1000sc100_100origin_data", s=1000, 
             sc=[100, 100])