使用 PIL Python 重新获取时，图像大小像素正在变化？

Question

fName = sys.argv[1]
X_original = Image.open(fName)

width, height = X_original.size

print(f'Original image  --> width:{width}, height:{height}')

scale = 0.90 

width_r, height_r = int(width*scale), int(height*scale)

print('Percentage of downsample:{:0.2f}%'.format(scale*100))

print(f'Resize image  --> width:{width_r}, height:{height_r}')

X_resize = X_original.resize((width_r, height_r))


X_original = np.array(X_original)
print('Original image shape (height, width) --> ', X_original.shape)

print('Total number of unique pixel values: ', len(np.unique(X_original)))

X_resize = np.array(X_resize)
print('Resize image shape (height, width) --> ', X_resize.shape)




print('Total number of unique pixel values: ', len(np.unique(X_resize)))

img_original = Image.fromarray(X_original)
img_original.save("img_original.png")

img_resize = Image.fromarray(X_resize)
img_resize.save("img_resize.png")

problem is:

**ouput ****

Total number of unique pixel values of orginal:15

Total number of unique pixel values after resize:50**

Answer 1

I think you're asking how/why additional colours/shades of grey are appearing in an image, despite making the size smaller.

I believe they're being created during the bicubic interpolation process.

See:

• https://en.wikipedia.org/wiki/Bicubic_interpolation
• https://pillow.readthedocs.io/en/stable/_modules/PIL/Image.html#Image.resize

If the image mode specifies a number of bits, such as "I;16", then the default filter is:py:data: PIL.Image.Resampling.NEAREST . Otherwise, the default filter is:py:data: PIL.Image.Resampling.BICUBIC .