Short-block variable-length trellis quantization

Tomas Eriksson; Mirek Novak; John B Anderson

doi:10.1109/DCC.2005.81

Short-block variable-length trellis quantization

Tomas Eriksson, Mirek Novak, John B Anderson

Department of Electrical and Information Technology

Research output: Chapter in Book/Report/Conference proceeding › Paper in conference proceeding › peer-review

1 Downloads (Pure)

Abstract

We present two methods for variable-rate trellis quantization. Both methods utilize trellis codes based on linear congruential (LC) recursions. LC code trellises have good pseudo-random properties and are easily adapted to serve reconstruction alphabets of different sizes. The first method finds an entropy-constrained code only by optimizing over a scale factor. The scale factor
modifies an initial reproducer alphabet in order to skew the associated set of codeword lengths. Using a Lagrangian formulation and the maximum a posteriori
(MAP) heuristic, we also develop an entropy-constrained trellis quantizer suitable for short blocks of data. Here the tailbiting BCJR algorithm is used to find the MAP path in the trellis. Simulation results for the Gaussian and Laplacian distributions show that the proposed method is competitive with the best in the literature.

Original language	English
Title of host publication	Proc., 2005 IEEE Data Compression Conf., Snowbird
Publisher	IEEE - Institute of Electrical and Electronics Engineers Inc.
Pages	251-260
Number of pages	10
ISBN (Print)	0-7695-2309-9
DOIs	https://doi.org/10.1109/DCC.2005.81
Publication status	Published - 2005
Event	2005 IEEE Data Compression Conf. - Snowbird, UT (USA) Duration: 2005 Mar 29 → 2005 Mar 31

Publication series

Name
ISSN (Print)	1068-0314

Conference

Conference	2005 IEEE Data Compression Conf.
Period	2005/03/29 → 2005/03/31

Subject classification (UKÄ)

Electrical Engineering, Electronic Engineering, Information Engineering

Free keywords

Data compression
trellis codes
BCJR algorithm
maximum a posteriori probability
rate-distortion theory

Access to Document

10.1109/DCC.2005.81

Cite this

@inproceedings{0e1e4764de9e4c07bc53e167ab76bf6d,

title = "Short-block variable-length trellis quantization",

abstract = "We present two methods for variable-rate trellis quantization. Both methods utilize trellis codes based on linear congruential (LC) recursions. LC code trellises have good pseudo-random properties and are easily adapted to serve reconstruction alphabets of different sizes. The first method finds an entropy-constrained code only by optimizing over a scale factor. The scale factor modifies an initial reproducer alphabet in order to skew the associated set of codeword lengths. Using a Lagrangian formulation and the maximum a posteriori (MAP) heuristic, we also develop an entropy-constrained trellis quantizer suitable for short blocks of data. Here the tailbiting BCJR algorithm is used to find the MAP path in the trellis. Simulation results for the Gaussian and Laplacian distributions show that the proposed method is competitive with the best in the literature.",

keywords = "Data compression, trellis codes, BCJR algorithm, maximum a posteriori probability, rate-distortion theory",

author = "Tomas Eriksson and Mirek Novak and Anderson, {John B}",

year = "2005",

doi = "10.1109/DCC.2005.81",

language = "English",

isbn = "0-7695-2309-9",

publisher = "IEEE - Institute of Electrical and Electronics Engineers Inc.",

pages = "251--260",

booktitle = "Proc., 2005 IEEE Data Compression Conf., Snowbird",

address = "United States",

note = "2005 IEEE Data Compression Conf. ; Conference date: 29-03-2005 Through 31-03-2005",

}

TY - GEN

T1 - Short-block variable-length trellis quantization

AU - Eriksson, Tomas

AU - Novak, Mirek

AU - Anderson, John B

PY - 2005

Y1 - 2005

N2 - We present two methods for variable-rate trellis quantization. Both methods utilize trellis codes based on linear congruential (LC) recursions. LC code trellises have good pseudo-random properties and are easily adapted to serve reconstruction alphabets of different sizes. The first method finds an entropy-constrained code only by optimizing over a scale factor. The scale factor modifies an initial reproducer alphabet in order to skew the associated set of codeword lengths. Using a Lagrangian formulation and the maximum a posteriori (MAP) heuristic, we also develop an entropy-constrained trellis quantizer suitable for short blocks of data. Here the tailbiting BCJR algorithm is used to find the MAP path in the trellis. Simulation results for the Gaussian and Laplacian distributions show that the proposed method is competitive with the best in the literature.

AB - We present two methods for variable-rate trellis quantization. Both methods utilize trellis codes based on linear congruential (LC) recursions. LC code trellises have good pseudo-random properties and are easily adapted to serve reconstruction alphabets of different sizes. The first method finds an entropy-constrained code only by optimizing over a scale factor. The scale factor modifies an initial reproducer alphabet in order to skew the associated set of codeword lengths. Using a Lagrangian formulation and the maximum a posteriori (MAP) heuristic, we also develop an entropy-constrained trellis quantizer suitable for short blocks of data. Here the tailbiting BCJR algorithm is used to find the MAP path in the trellis. Simulation results for the Gaussian and Laplacian distributions show that the proposed method is competitive with the best in the literature.

KW - Data compression

KW - trellis codes

KW - BCJR algorithm

KW - maximum a posteriori probability

KW - rate-distortion theory

U2 - 10.1109/DCC.2005.81

DO - 10.1109/DCC.2005.81

M3 - Paper in conference proceeding

SN - 0-7695-2309-9

SP - 251

EP - 260

BT - Proc., 2005 IEEE Data Compression Conf., Snowbird

PB - IEEE - Institute of Electrical and Electronics Engineers Inc.

T2 - 2005 IEEE Data Compression Conf.

Y2 - 29 March 2005 through 31 March 2005

ER -

Short-block variable-length trellis quantization

Abstract

Publication series

Conference

Subject classification (UKÄ)

Free keywords

Access to Document

Fingerprint

Cite this