|
Published Articles >> Table of Contents >> Abstract
Data Compression Conference (DCC '04)
p. 52
Minimax Multiresolution Scalar Quantization
Nima Sarshar, McMaster University, Hamilton, Canada
Xiaolin Wu, Polytechnic University, Brooklyn, New York
Full Article Text:
  
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/DCC.2004.1281450
Send link to a friend
| Abstract |
|
We consider the problem of design and analysis of optimal L\infty (minmax) multi-resolution scalar quantizers (MRSQ). The overall multi-resolution L\infty distortion of an MRSQ is defined to be a weighted sum of L\infty distortions over all refinement levels of the MRSQ. The weight for a refinement level usually denotes the probability that the MRSQ will operate at that level (rate). An interesting relation of the problem to the design of optimal binary prefix codes under a code cell contiguity constraint is established. Lower bounds for the overall multi-resolution L\infty distortion are derived based on this relation. Provably optimal as well as fast, near optimal algorithms are also developed for practically interesting scenarios. Furthermore, the performance penalty incurred by making a scalable quantizer embedded (progressively refinable) is analyzed. It is shown that constraining the quantizers to be embedded would on average increase the L\infty quantization error by at least 44%.
|
 Additional Information
|
Index Terms- Multiresolution signal representation, scalar quantization, minimax
optimization, progressive codes
Citation:
Nima Sarshar, Xiaolin Wu,
"Minimax Multiresolution Scalar Quantization,"
dcc,
p. 52,
Data Compression Conference (DCC '04),
2004
|
|
