15. Image quantize
15.1. Quantize
Use theImage.quantize(colors=256, method=None, kmeans=0, palette=None, dither=Dither.FLOYDSTEINBERG) to return an image converted to P mode with the specified number of colors.
colors - The desired number of colors, <= 256
method - Quantize.MEDIANCUT (median cut), Quantize.MAXCOVERAGE (maximum coverage), Quantize.FASTOCTREE (fast octree), Quantize.LIBIMAGEQUANT (libimagequant; check support using PIL.features.check_feature() with feature=”libimagequant”).
By default, Quantize.MEDIANCUT will be used.
The exception to this is RGBA images. Quantize.MEDIANCUT and Quantize.MAXCOVERAGE do not support RGBA images, so Quantize.FASTOCTREE is used by default instead.
kmeans - Integer
palette - Quantize to the palette of a given Image.
dither - Dithering method, used when converting from mode “RGB” to “P” or from “RGB” or “L” to “1”. Available methods are Dither.NONE or Dither.FLOYDSTEINBERG (default).
15.2. Quantize number of colours
The code below produces a series of images converted to P mode with the specified number of colors.
from PIL import Image
with Image.open("test_images/lorne.jpg") as im:
im1 = im.quantize(colors=4)
im1.save("Image/image_quantize4.png")
im1 = im.quantize(colors=8)
im1.save("Image/image_quantize8.png")
im1 = im.quantize(colors=32)
im1.save("Image/image_quantize32.png")
15.3. Quantize by image palette
The code below produces an image converted to P mode with the specified image palette.
The code below.
from PIL import Image
with Image.open("new_images/draw_palette_4.png") as im:
im_pal1 = im.convert(mode="P", palette=Image.Palette.WEB)
with Image.open("new_images/draw_palette5.png") as im:
im_pal2 = im.convert(mode="P", palette=Image.Palette.ADAPTIVE)
with Image.open("test_images/lorne.jpg") as im:
im1 = im.quantize(palette=im_pal1)
im1.save("Image/image_quantize_im_pal.png")
im2 = im.quantize(palette=im_pal2)
im2.save("Image/image_quantize_im_pal2.png")
