Unsupervised Learning Approach to Adaptive Differential Pulse Code Modulation

Norman C. Griswold, Khalid Sayood

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

This research is concerned with investigating the problem of data compression utilizing an unsupervised estimation algorithm. This extends previous work utilizing a hybrid source coder which combines an orthogonal transformation with differential pulse code modulation (DPCM). The data compression is achieved in the DPCM loop, and it is the quantizer of this scheme which is approached from an unsupervised learning procedure. The distribution defining the quantizer is represented as a set of separable Laplacian mixture densities for two-dimensional images. The condition of idcntifiability is shown for the Laplacian case and a decision directed estimate of both the active distribution parameters and the mixing parameters are discussed in view of a Bayesian structure. The decision directed estimators, although not optimum, provide a realizable structure for estimating the parameters which define a distribution which has become active. These parameters are then used to scale the optimum (in the mean square error sense) Laplacian quantizer. The decision criteria is modified to prevent convergence to a single distribution which in effect is the default condition for a variance estimator. This investigation was applied to a test image and the resulting data demonstrate improvement over other techniques using fixed bit assignments and ideal channel conditions.

Original languageEnglish (US)
Pages (from-to)380-391
Number of pages12
JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
VolumePAMI-4
Issue number4
DOIs
StatePublished - Jan 1 1982

Fingerprint

Differential pulse code modulation
Unsupervised learning
Unsupervised Learning
Data compression
Modulation
Data Compression
Mean square error
Orthogonal Transformation
Variance Estimator
Estimation Algorithms
Assignment
Estimator
Estimate
Demonstrate

Keywords

  • Adaptive differential pulse code modulation (DPCM)
  • adaptive quantization
  • bandwidth compression
  • decision directed estimates
  • unsupervised learning

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition
  • Computational Theory and Mathematics
  • Artificial Intelligence
  • Applied Mathematics

Cite this

Unsupervised Learning Approach to Adaptive Differential Pulse Code Modulation. / Griswold, Norman C.; Sayood, Khalid.

In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. PAMI-4, No. 4, 01.01.1982, p. 380-391.

Research output: Contribution to journalArticle

@article{2cfc2cf701fc40098218e4e956e1ec79,
title = "Unsupervised Learning Approach to Adaptive Differential Pulse Code Modulation",
abstract = "This research is concerned with investigating the problem of data compression utilizing an unsupervised estimation algorithm. This extends previous work utilizing a hybrid source coder which combines an orthogonal transformation with differential pulse code modulation (DPCM). The data compression is achieved in the DPCM loop, and it is the quantizer of this scheme which is approached from an unsupervised learning procedure. The distribution defining the quantizer is represented as a set of separable Laplacian mixture densities for two-dimensional images. The condition of idcntifiability is shown for the Laplacian case and a decision directed estimate of both the active distribution parameters and the mixing parameters are discussed in view of a Bayesian structure. The decision directed estimators, although not optimum, provide a realizable structure for estimating the parameters which define a distribution which has become active. These parameters are then used to scale the optimum (in the mean square error sense) Laplacian quantizer. The decision criteria is modified to prevent convergence to a single distribution which in effect is the default condition for a variance estimator. This investigation was applied to a test image and the resulting data demonstrate improvement over other techniques using fixed bit assignments and ideal channel conditions.",
keywords = "Adaptive differential pulse code modulation (DPCM), adaptive quantization, bandwidth compression, decision directed estimates, unsupervised learning",
author = "Griswold, {Norman C.} and Khalid Sayood",
year = "1982",
month = "1",
day = "1",
doi = "10.1109/TPAMI.1982.4767269",
language = "English (US)",
volume = "PAMI-4",
pages = "380--391",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
issn = "0162-8828",
publisher = "IEEE Computer Society",
number = "4",

}

TY - JOUR

T1 - Unsupervised Learning Approach to Adaptive Differential Pulse Code Modulation

AU - Griswold, Norman C.

AU - Sayood, Khalid

PY - 1982/1/1

Y1 - 1982/1/1

N2 - This research is concerned with investigating the problem of data compression utilizing an unsupervised estimation algorithm. This extends previous work utilizing a hybrid source coder which combines an orthogonal transformation with differential pulse code modulation (DPCM). The data compression is achieved in the DPCM loop, and it is the quantizer of this scheme which is approached from an unsupervised learning procedure. The distribution defining the quantizer is represented as a set of separable Laplacian mixture densities for two-dimensional images. The condition of idcntifiability is shown for the Laplacian case and a decision directed estimate of both the active distribution parameters and the mixing parameters are discussed in view of a Bayesian structure. The decision directed estimators, although not optimum, provide a realizable structure for estimating the parameters which define a distribution which has become active. These parameters are then used to scale the optimum (in the mean square error sense) Laplacian quantizer. The decision criteria is modified to prevent convergence to a single distribution which in effect is the default condition for a variance estimator. This investigation was applied to a test image and the resulting data demonstrate improvement over other techniques using fixed bit assignments and ideal channel conditions.

AB - This research is concerned with investigating the problem of data compression utilizing an unsupervised estimation algorithm. This extends previous work utilizing a hybrid source coder which combines an orthogonal transformation with differential pulse code modulation (DPCM). The data compression is achieved in the DPCM loop, and it is the quantizer of this scheme which is approached from an unsupervised learning procedure. The distribution defining the quantizer is represented as a set of separable Laplacian mixture densities for two-dimensional images. The condition of idcntifiability is shown for the Laplacian case and a decision directed estimate of both the active distribution parameters and the mixing parameters are discussed in view of a Bayesian structure. The decision directed estimators, although not optimum, provide a realizable structure for estimating the parameters which define a distribution which has become active. These parameters are then used to scale the optimum (in the mean square error sense) Laplacian quantizer. The decision criteria is modified to prevent convergence to a single distribution which in effect is the default condition for a variance estimator. This investigation was applied to a test image and the resulting data demonstrate improvement over other techniques using fixed bit assignments and ideal channel conditions.

KW - Adaptive differential pulse code modulation (DPCM)

KW - adaptive quantization

KW - bandwidth compression

KW - decision directed estimates

KW - unsupervised learning

UR - http://www.scopus.com/inward/record.url?scp=0020151722&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0020151722&partnerID=8YFLogxK

U2 - 10.1109/TPAMI.1982.4767269

DO - 10.1109/TPAMI.1982.4767269

M3 - Article

VL - PAMI-4

SP - 380

EP - 391

JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

SN - 0162-8828

IS - 4

ER -