Down-sampling and anti-aliasing
As we have seen, down-sampling is not very good for shrinking images as it creates an aliasing effect. For instance, if we try to resize (down-sample) the original image by reducing the width and height a factor of 5, we shall get such patchy and bad output.
Anti-aliasing
The problem here is that a single pixel in the output image corresponds to 25 pixels in the input image, but we are sampling the value of a single pixel instead. We should be averaging over a small area in the input image. This can be done using ANTIALIAS (a high-quality down-sampling filter); this is how you can do it:
im = im.resize((im.width//5, im.height//5), Image.ANTIALIAS)
pylab.figure(figsize=(15,10)), pylab.imshow(im), pylab.show()
An image, the same as the previous one but with much nicer quality (almost without any artifact/aliasing effect), is created using PIL with anti-aliasing:
Anti-aliasing is generally done by smoothing an image (via convolution of the image with a low-pass filter such as a Gaussian filter) before down-sampling.
Let's now use the scikit-image transform module's rescale() function with anti-aliasing to overcome the aliasing problem for another image, namely the umbc.png image:
im = imread('../images/umbc.png')
im1 = im.copy()
pylab.figure(figsize=(20,15))
for i in range(4):
pylab.subplot(2,2,i+1), pylab.imshow(im1, cmap='gray'), pylab.axis('off')
pylab.title('image size = ' + str(im1.shape[1]) + 'x' + str(im1.shape[0]))
im1 = rescale(im1, scale = 0.5, multichannel=True, anti_aliasing=False)
pylab.subplots_adjust(wspace=0.1, hspace=0.1)
pylab.show()
The next screenshot shows the output of the previous code. As can be seen, the image is down-sampled to create smaller and smaller output—the aliasing effect becomes more prominent when no anti-aliasing technique is used:
Let's change the line of code to use anti-aliasing:
im1 = rescale(im1, scale = 0.5, multichannel=True, anti_aliasing=True)
This yields an image of better quality:
