2004
DOI: 10.1109/tit.2004.831787
|View full text |Cite
|
Sign up to set email alerts
|

Suboptimality of the Karhunen–LoÈve Transform for Transform Coding

Abstract: Abstract-We examine the performance of the Karhunen-Loève transform (KLT) for transform coding applications. The KLT has long been viewed as the best available block transform for a system that orthogonally transforms a vector source, scalar quantizes the components of the transformed vector using optimal bit allocation, and then inverse transforms the vector. This paper treats fixed-rate and variable-rate transform codes of non-Gaussian sources. The fixed-rate approach uses an optimal fixed-rate scalar quanti… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

0
12
0

Year Published

2008
2008
2021
2021

Publication Types

Select...
4
3
3

Relationship

0
10

Authors

Journals

citations
Cited by 67 publications
(12 citation statements)
references
References 11 publications
0
12
0
Order By: Relevance
“…The use of DCT in data compression and pattern classification has been increased in recent years, mainly due to the fact its performance is much closer to the results obtained by the Karhunen-Loève transform which is considered optimal for a variety of criteria such as mean square error of truncation and entropy [4], [5]. This paper demonstrates the potential of DCT, CPSO and GMM in speech recognition and the viability of them [6].…”
Section: Introductionmentioning
confidence: 74%
“…The use of DCT in data compression and pattern classification has been increased in recent years, mainly due to the fact its performance is much closer to the results obtained by the Karhunen-Loève transform which is considered optimal for a variety of criteria such as mean square error of truncation and entropy [4], [5]. This paper demonstrates the potential of DCT, CPSO and GMM in speech recognition and the viability of them [6].…”
Section: Introductionmentioning
confidence: 74%
“…5. Since x follows a Gaussian mixture distribution, the KLT obtained from the covariance of x (which implicitly performs a second-order approximation of the distribution) is suboptimal in MSE sense [50]. Especially, with many possible image edge locations and different orientations, the underlying distribution may contain a large number of mixtures (i.e., a large M ), which makes learning a model from average statistics inefficient.…”
Section: B Theoretical Justification For Ea-gbtsmentioning
confidence: 99%
“…We can compare it with algorithms based on the unique properties of wavelets and fractals as alternative coding methods [13]. The technique described in this paper has the lossy coder of the spectral coefficients of the Karhunen-Loève transform [11,5]. It seems to be a better solution.…”
Section: Lossy Astronomical Image Compressionmentioning
confidence: 99%