Timeline for What properties make certain wavelets "better" than others in image compression?
Current License: CC BY-SA 3.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 11, 2014 at 16:37 | comment | added | Lutz Lehmann | Nice article. The information in the wikipedia article corresponds to the sources cited, essentially Daubechies "10 Lectures", so it may be outdated with respect to JPEG2000. One correction: biorthogonal is not redundant. The biorthogonality conditions impose exactly inverse filter banks. Redundant transformantions start with framelets. | |
| Feb 7, 2013 at 17:02 | comment | added | Spacey | @datageist Fantastic answer! Also, another reason that 9/7 came to exist in the first place was because it was an alternative way to factor the reconstruction polynomial, with the constraint that the filters be symmetric. This way, the phase response remains linear. (In contrast, a daub4 wavelet, while an FIR, is asymmetric and induces non-linear phases in a processed signal). 9/7 was used in JPEG because of the subjective inclination for us to like linear over non-linear distortions in images. | |
| Jul 6, 2012 at 18:02 | history | answered | datageist♦ | CC BY-SA 3.0 |