Embedded zerotrees of wavelet transforms: Difference between revisions

Content deleted Content added
Added some explanation of how EZW works. Could probably do with a diagram or two, but I don't have anything suitable to hand.
m Add ref to subband coding
Line 1:
'''EZW''' ('''Embedded Zerotree'''s of '''Wavelet''' Transforms) is a lossy [[Image compression|image compression]] [[algorithm]]. At low bit rates
(i.e. high compression ratios) most of the coefficients produced by a [[Sub-band coding|subband transform]] (such as the [[Wavelet transform|wavelet transform]])
will be zero, or very close to zero. This occurs because "real world" images tend to contain mostly low frequency information (highly correlated). However where high frequency information does occur (such as edges in the image) this is particularly important in terms of human perception of the image quality, and thus must be represented accurately in any high quality coding scheme.