ColorJitter in PyTorch-Python Tutorial-php.cn

Home

Backend Development

Python Tutorial

ColorJitter in PyTorch

Patricia Arquette

Dec 30, 2024 am 09:20 AM

Buy Me a Coffee☕

ColorJitter() can change the brightness, contrast, saturation and hue of zero or more images as shown below:

*Memos:

The 1st argument for initialization is brightness(Optional-Default:0-Type:float or tuple/list(float)): *Memos:
- It's the range of the brightness [min, max].
- It must be 0
- A single value is converted to [max(0, 1-brightness), 1 brightness].
- A tuple or list must be the 1D with 2 elements. *The 1st element must be less than or equal to the 2nd element.
The 2nd argument for initialization is contrast(Optional-Default:0-Type:float or tuple/list(float)): *Memos:
- It's the range of the contrast [min, max].
- It must be 0
- A single value is converted to [max(0, 1-contrast), 1 contrast].
- A tuple or list must be the 1D with 2 elements. *The 1st element must be less than or equal to the 2nd element.
The 3rd argument for initialization is saturation(Optional-Default:0-Type:float or tuple/list(float)): *Memos:
- It's the range of the saturation [min, max].
- It must be 0
- A single value is converted to [max(0, 1-saturation), 1 saturation].
- A tuple or list must be the 1D with 2 elements. *The 1st element must be less than or equal to the 2nd element.
The 4th argument for initialization is hue(Optional-Default:0-Type:float or tuple/list(float)): *Memos:
- It's the range of the hue [min, max].
- It must be -0.5
- A single value is converted to [-hue, hue].
- A tuple or list must be the 1D with 2 elements. *The 1st element must be less than or equal to the 2nd element.
The 1st argument is img(Required-Type:PIL Image or tensor/tuple/list(int or float)): *Memos:
- It must be 2D or 3D. For 3D, the deepest D must have one element.
- 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 ColorJitter

colorjitter = ColorJitter()
colorjitter = ColorJitter(brightness=0,
                          contrast=0,
                          saturation=0,
                          hue=0)
colorjitter = ColorJitter(brightness=(1.0, 2.0),
                          contrast=(1.0, 1.0),
                          saturation=(1.0, 1.0),
                          hue=(0.0, 0.0))
colorjitter
# ColorJitter()

print(colorjitter.brightness)
# None

print(colorjitter.contrast)
# None

print(colorjitter.saturation)
# None

print(colorjitter.hue)
# None

origin_data = OxfordIIITPet(
    root="data",
    transform=None
    # transform=ColorJitter()
    # colorjitter = ColorJitter(brightness=0,
    #                           contrast=0,
    #                           saturation=0,
    #                           hue=0)
    # transform=ColorJitter(brightness=(1.0, 1.0),
    #                       contrast=(1.0, 1.0),
    #                       saturation=(1.0, 1.0),
    #                       hue=(0.0, 0.0))
)

p2bright_data = OxfordIIITPet( # `p` is plus.
    root="data",
    transform=ColorJitter(brightness=2.0)
    # transform=ColorJitter(brightness=(0.0, 3.0))
)

p2p2bright_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(brightness=(2.0, 2.0))
)

p05p05bright_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(brightness=(0.5, 0.5))
)

p2contra_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(contrast=2.0)
    # transform=ColorJitter(contrast=(0.0, 3.0))
)

p2p2contra_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(contrast=(2.0, 2.0))
)

p05p05contra_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(contrast=(0.5, 0.5))
)

p2satura_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(saturation=2.0)
    # transform=ColorJitter(saturation=(0.0, 3.0))
)

p2p2satura_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(saturation=(2.0, 2.0))
)

p05p05satura_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(saturation=(0.5, 0.5))
)

p05hue_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(hue=0.5)
    # transform=ColorJitter(hue=(-0.5, 0.5))
)

p025p025hue_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(hue=(0.25, 0.25))
)

m025m025hue_data = OxfordIIITPet( # `m` is minus.
    root="data",
    transform=ColorJitter(hue=(-0.25, -0.25))
)

import matplotlib.pyplot as plt

def show_images(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_images(data=origin_data, main_title="origin_data")
show_images(data=p2bright_data, main_title="p2bright_data")
show_images(data=p2p2bright_data, main_title="p2p2bright_data")
show_images(data=p05p05bright_data, main_title="p05p05bright_data")

show_images(data=origin_data, main_title="origin_data")
show_images(data=p2contra_data, main_title="p2contra_data")
show_images(data=p2p2contra_data, main_title="p2p2contra_data")
show_images(data=p05p05contra_data, main_title="p05p05contra_data")

show_images(data=origin_data, main_title="origin_data")
show_images(data=p2satura_data, main_title="p2satura_data")
show_images(data=p2p2satura_data, main_title="p2p2satura_data")
show_images(data=p05p05satura_data, main_title="p05p05satura_data")

show_images(data=origin_data, main_title="origin_data")
show_images(data=p05hue_data, main_title="p05hue_data")
show_images(data=p025p025hue_data, main_title="p025p025hue_data")
show_images(data=m025m025hue_data, main_title="m025m025hue_data")

ColorJitter in PyTorch

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

colorjitter = ColorJitter()
colorjitter = ColorJitter(brightness=0,
                          contrast=0,
                          saturation=0,
                          hue=0)
colorjitter = ColorJitter(brightness=(1.0, 2.0),
                          contrast=(1.0, 1.0),
                          saturation=(1.0, 1.0),
                          hue=(0.0, 0.0))
colorjitter
# ColorJitter()

print(colorjitter.brightness)
# None

print(colorjitter.contrast)
# None

print(colorjitter.saturation)
# None

print(colorjitter.hue)
# None

origin_data = OxfordIIITPet(
    root="data",
    transform=None
    # transform=ColorJitter()
    # colorjitter = ColorJitter(brightness=0,
    #                           contrast=0,
    #                           saturation=0,
    #                           hue=0)
    # transform=ColorJitter(brightness=(1.0, 1.0),
    #                       contrast=(1.0, 1.0),
    #                       saturation=(1.0, 1.0),
    #                       hue=(0.0, 0.0))
)

p2bright_data = OxfordIIITPet( # `p` is plus.
    root="data",
    transform=ColorJitter(brightness=2.0)
    # transform=ColorJitter(brightness=(0.0, 3.0))
)

p2p2bright_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(brightness=(2.0, 2.0))
)

p05p05bright_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(brightness=(0.5, 0.5))
)

p2contra_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(contrast=2.0)
    # transform=ColorJitter(contrast=(0.0, 3.0))
)

p2p2contra_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(contrast=(2.0, 2.0))
)

p05p05contra_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(contrast=(0.5, 0.5))
)

p2satura_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(saturation=2.0)
    # transform=ColorJitter(saturation=(0.0, 3.0))
)

p2p2satura_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(saturation=(2.0, 2.0))
)

p05p05satura_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(saturation=(0.5, 0.5))
)

p05hue_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(hue=0.5)
    # transform=ColorJitter(hue=(-0.5, 0.5))
)

p025p025hue_data = OxfordIIITPet(
    root="data",
    transform=ColorJitter(hue=(0.25, 0.25))
)

m025m025hue_data = OxfordIIITPet( # `m` is minus.
    root="data",
    transform=ColorJitter(hue=(-0.25, -0.25))
)

import matplotlib.pyplot as plt

def show_images(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_images(data=origin_data, main_title="origin_data")
show_images(data=p2bright_data, main_title="p2bright_data")
show_images(data=p2p2bright_data, main_title="p2p2bright_data")
show_images(data=p05p05bright_data, main_title="p05p05bright_data")

show_images(data=origin_data, main_title="origin_data")
show_images(data=p2contra_data, main_title="p2contra_data")
show_images(data=p2p2contra_data, main_title="p2p2contra_data")
show_images(data=p05p05contra_data, main_title="p05p05contra_data")

show_images(data=origin_data, main_title="origin_data")
show_images(data=p2satura_data, main_title="p2satura_data")
show_images(data=p2p2satura_data, main_title="p2p2satura_data")
show_images(data=p05p05satura_data, main_title="p05p05satura_data")

show_images(data=origin_data, main_title="origin_data")
show_images(data=p05hue_data, main_title="p05hue_data")
show_images(data=p025p025hue_data, main_title="p025p025hue_data")
show_images(data=m025m025hue_data, main_title="m025m025hue_data")

ColorJitter in PyTorch

The above is the detailed content of ColorJitter in PyTorch. For more information, please follow other related articles on the PHP Chinese website!

Statement

The content of this article is voluntarily contributed by netizens, and the copyright belongs to the original author. This site does not assume corresponding legal responsibility. If you find any content suspected of plagiarism or infringement, please contact admin@php.cn

Merging Lists in Python: Choosing the Right MethodMay 14, 2025 am 12:11 AM

TomergelistsinPython,youcanusethe operator,extendmethod,listcomprehension,oritertools.chain,eachwithspecificadvantages:1)The operatorissimplebutlessefficientforlargelists;2)extendismemory-efficientbutmodifiestheoriginallist;3)listcomprehensionoffersf

How to concatenate two lists in python 3?May 14, 2025 am 12:09 AM

In Python 3, two lists can be connected through a variety of methods: 1) Use operator, which is suitable for small lists, but is inefficient for large lists; 2) Use extend method, which is suitable for large lists, with high memory efficiency, but will modify the original list; 3) Use * operator, which is suitable for merging multiple lists, without modifying the original list; 4) Use itertools.chain, which is suitable for large data sets, with high memory efficiency.

Python concatenate list stringsMay 14, 2025 am 12:08 AM

Using the join() method is the most efficient way to connect strings from lists in Python. 1) Use the join() method to be efficient and easy to read. 2) The cycle uses operators inefficiently for large lists. 3) The combination of list comprehension and join() is suitable for scenarios that require conversion. 4) The reduce() method is suitable for other types of reductions, but is inefficient for string concatenation. The complete sentence ends.

Python execution, what is that?May 14, 2025 am 12:06 AM

PythonexecutionistheprocessoftransformingPythoncodeintoexecutableinstructions.1)Theinterpreterreadsthecode,convertingitintobytecode,whichthePythonVirtualMachine(PVM)executes.2)TheGlobalInterpreterLock(GIL)managesthreadexecution,potentiallylimitingmul

Python: what are the key featuresMay 14, 2025 am 12:02 AM

Key features of Python include: 1. The syntax is concise and easy to understand, suitable for beginners; 2. Dynamic type system, improving development speed; 3. Rich standard library, supporting multiple tasks; 4. Strong community and ecosystem, providing extensive support; 5. Interpretation, suitable for scripting and rapid prototyping; 6. Multi-paradigm support, suitable for various programming styles.

Python: compiler or Interpreter?May 13, 2025 am 12:10 AM

Python is an interpreted language, but it also includes the compilation process. 1) Python code is first compiled into bytecode. 2) Bytecode is interpreted and executed by Python virtual machine. 3) This hybrid mechanism makes Python both flexible and efficient, but not as fast as a fully compiled language.

Python For Loop vs While Loop: When to Use Which?May 13, 2025 am 12:07 AM

Useaforloopwheniteratingoverasequenceorforaspecificnumberoftimes;useawhileloopwhencontinuinguntilaconditionismet.Forloopsareidealforknownsequences,whilewhileloopssuitsituationswithundeterminediterations.

Python loops: The most common errorsMay 13, 2025 am 12:07 AM

Pythonloopscanleadtoerrorslikeinfiniteloops,modifyinglistsduringiteration,off-by-oneerrors,zero-indexingissues,andnestedloopinefficiencies.Toavoidthese:1)Use'i

See all articles