.point()

In Pillow, the .point() method is used to apply a function or lookup table to each pixel in an image. It is useful for performing operations like thresholding, gamma correction, or channel manipulation.

Syntax

Image.point(lut, mode=None)

Parameters:

• lut: This parameter can be:
  • A lookup table: A list or sequence with 256 values for 8-bit images per band (e.g., 768 values for RGB). For 16-bit images, 65536 values are required per band.
  • A function: Takes a single integer (0–255) and returns a value. It’s called once for each possible pixel value to build a lookup table internally.
• mode (optional, str): The mode of the output image. Use this if you want to change the image type during transformation. Common modes include:
  • "L": 8-bit pixels, black and white (grayscale)
  • "RGB": 3x8-bit pixels, true color
  • "RGBA": 4x8-bit pixels, true color with alpha channel
  • "1": 1-bit pixels, black and white
  • "P": 8-bit pixels, uses a color palette to map to other modes

Note: Most uses of .point() don’t require the argument mode unless there is a need to explicitly change the image type (e.g., converting grayscale to binary).

Return value:

This method returns a new Image object with the transformed pixel data.

Example

This example creates a horizontal grayscale gradient, inverts it using .point(), and saves the result:

from PIL import Image

# Create a horizontal grayscale gradient image
width, height = 256, 50
image = Image.new("L", (width, height))
for x in range(width):
  for y in range(height):
    image.putpixel((x, y), x)

# Invert grayscale values using .point()
inverted = image.point(lambda p: 255 - p)

# Save the result
inverted.save("inverted-gradient.png")

Copy to clipboard
Copy to clipboard

The output for the example will be: