1993
DOI: 10.1109/18.212286
|View full text |Cite
|
Sign up to set email alerts
|

Vector quantizers with direct sum codebooks

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
30
0

Year Published

1994
1994
2003
2003

Publication Types

Select...
6
2

Relationship

0
8

Authors

Journals

citations
Cited by 81 publications
(30 citation statements)
references
References 12 publications
0
30
0
Order By: Relevance
“…The generalised decoder uses a direct-sum structure that is jointly optimized for the source distribution and the uniform encoder. The direct-sum algorithm has been used successfully in the design of residual VQ Barnes and Frost (1993). Barnes and Holder (1993) reported significant improvements over a uniform decoder, especially at low bit rates.…”
Section: Successive Approximation Methodsmentioning
confidence: 98%
See 1 more Smart Citation
“…The generalised decoder uses a direct-sum structure that is jointly optimized for the source distribution and the uniform encoder. The direct-sum algorithm has been used successfully in the design of residual VQ Barnes and Frost (1993). Barnes and Holder (1993) reported significant improvements over a uniform decoder, especially at low bit rates.…”
Section: Successive Approximation Methodsmentioning
confidence: 98%
“…The design of a residual coder considers the type of coder to use and the rate assigned for each stage. The coders should be jointly optimized (Barnes and Frost, 1993), although in practice this is not always possible owing to the computational complexity. Multistage residual coders are associated with the following advantages:…”
Section: Multistage Residual Quantization Methodsmentioning
confidence: 99%
“…The procedure is initialized with a set of P stage codebooks that are typically obtained by traditional sequential design [9]. It then iterates over the stages and optimizes each stage codebook while keeping the remaining stages fixed [11]. The re-optimization of quantizer Q p consists of updating its codebook.…”
Section: A Overview Of Msvq and Ec-msvq Designmentioning
confidence: 99%
“…Second, the quantizer codebooks are optimized for the current training set partitions in what is termed decoder optimization. This is achieved by iterating over the stages and optimizing each stage codebook while keeping the remaining stages fixed [29]. The reoptimization of quantizer consists of updating its codebook.…”
Section: A Overviewmentioning
confidence: 99%