Invert the Colors of an Image Randomly with a given Probability in PyTorch

In this article, we will discuss how to invert the colors of an image randomly with a given probability in Python using PyTorch.

RandomInvert() method

RandomInvert() method accepts PIL and tensor image as input. With tensor, we provide shapes in [3, H, W], where H and W are the height and width of the image. This method returns a Randomly colored inverted image. The below syntax is used to invert the colors of an image randomly with a given probability.

Syntax: torchvision.transforms.RandomInvert(p)(img)

Parameter:

• img: This is our input image to be inverted.
• p: it’s probability in the range of (0,1). 

Return: This method returns a Randomly color inverted image. if p=0 then it will return the original image and if p=1 then it will return color inverted image.

The below image is used for demonstration:

Example:

The following program is to understand how to invert the colors of an image randomly with a given probability.

Python3

# import required libraries
import torch
import torchvision.transforms as T
from PIL import Image

# read image from your computer
img = Image.open('a.png')

# invert the colors with probability=1
transform = T.RandomInvert(p=1)
img = transform(img)

# display result
img.show()

Output: