WAVELET BASED DECOMPOSITION OF NATURAL IMAGES IN INDEPENDENT RESOLUTION LEVELS

Antonio Turiel, N'estor Parga
Antonio.Turiel@lps.ens.fr, parga@delta.ft.uam.es

Efficient coding of natural scenes requires the use of a proper filter to reduce their redundancy. In this work we show that nongaussian statistical properties of nat­ ural scenes uniquely define a wavelet filter that decom­ poses the image in a set of statistically independent res­ olution levels. The spatial statistical dependences still present at a fixed scale are extremely short­ranged.