Title :
Mathematical properties of the JPEG2000 wavelet filters
Author :
Unser, Michael ; Blu, Thierry
Author_Institution :
Biomed. Imaging Group, Swiss Fed. Inst. of Technol. Lausanne, Switzerland
Abstract :
The LeGall 5/3 and Daubechies 9/7 filters have risen to special prominence because they were selected for inclusion in the JPEG2000 standard. We determine their key mathematical features: Riesz bounds, order of approximation, and regularity (Holder and Sobolev). We give approximation theoretic quantities such as the asymptotic constant for the L2 error and the angle between the analysis and synthesis spaces which characterizes the loss of performance with respect to an orthogonal projection. We also derive new asymptotic error formulae that exhibit bound constants that are proportional to the magnitude of the first nonvanishing moment of the wavelet. The Daubechies 9/7 stands out because it is very close to orthonormal, but this turns out to be slightly detrimental to its asymptotic performance when compared to other wavelets with four vanishing moments.
Keywords :
approximation theory; data compression; filtering theory; image coding; wavelet transforms; Daubechies filters; Holder regularity; JPEG2000 wavelet filters; LeGall filters; Riesz bounds; Sobolev regularity; approximation order; approximation theory; orthogonal projection; still digital picture compression; Decoding; Discrete cosine transforms; Filter bank; Image coding; Image quality; Performance analysis; Performance loss; Spline; Transform coding; Wavelet transforms;
Journal_Title :
Image Processing, IEEE Transactions on
DOI :
10.1109/TIP.2003.812329