Machine Learning for OpenCV 4 - Second Edition

Compressing color spaces using k-means

One interesting use case of k-means is the compression of image color spaces. For example, a standard color image comes with 24-bit color depth, providing for a total of 16,777,216 color varieties. However, in most images, a large number of colors will be unused, and many of the pixels in the image will have similar values. The compressed image can then be sent over the internet at a faster speed, and at the receiver end, it can be decompressed to get back the original image. Hence, reducing the storage and transmission cost. However, the image color space compression will be lossy and you might not notice fine details in the image after compression.

Alternatively, we can use k-means to reduce the ...