Abstract
Several lifting-based filter banks (LBFBs) are proposed for lossless-to-lossy image coding. However, the conventional methods can not achieve good coding performance due to many rounding operations, and they may have wide dynamic range of lifting coefficients due to inverse matrix. This paper presents a more practical lifting factorization of M-channel perfect reconstruction filter bank (PRFB). Using the block parallel system of PRFBs with an imposed condition, we can factorize them into a novel lifting structure which is called the parallel lifting. The validity of our structures is shown by applying them to lossless-to-lossy image coding.
| Original language | English |
|---|---|
| Pages (from-to) | 2245-2248 |
| Number of pages | 4 |
| Journal | European Signal Processing Conference |
| Publication status | Published - 2009 Dec 1 |
| Event | 17th European Signal Processing Conference, EUSIPCO 2009 - Glasgow, United Kingdom Duration: 2009 Aug 24 → 2009 Aug 28 |
Keywords
- Block parallel system
- Lossless-to-lossy image coding
- Parallel lifting
- Perfect reconstruction filter bank (PRFB)
ASJC Scopus subject areas
- Signal Processing
- Electrical and Electronic Engineering