Abstract/Details

ADAPTIVE SOURCE MODELS FOR DATA COMPRESSION

RAMABADRAN, TENKASI V. 
 University of Notre Dame ProQuest Dissertations Publishing,  1987. 8712954.

Abstract (summary)

Noiseless compression of finite sequences can be viewed as a two-step process consisting of (i) source modeling and (ii) coding. Efficient coding techniques are now known. Therefore, the major problem in noiseless data compression is one of building a source model of a given complexity, which is well matched to the sequence to be compressed.

In this dissertation, a generalized dependent source model called the Conditioned Source Model is studied. The concepts of splitting and merging of contexts (i.e. conditioning events) are introduced and some simple and useful properties of the model are derived using well known Information Theory inequalities. The problem of selection of a good set of contexts for the Conditioned Source Model is then addressed and an approach to adaptive source modeling with selective context splitting is proposed. This method is more practical with a smaller alphabet size and so a technique for alphabet reduction with statistics separation is discussed next.

Finally, a new data compression algorithm, called CRAM, is presented. CRAM uses the alphabet reduction technique, builds the source model adaptively with selective context splitting, and encodes a given sequence by means of arithmetic coding. It is shown to be effective for a variety of computer files.

Indexing (details)


Subject
Electrical engineering
Classification
0544: Electrical engineering
Identifier / keyword
Applied sciences
Title
ADAPTIVE SOURCE MODELS FOR DATA COMPRESSION
Author
RAMABADRAN, TENKASI V.
Number of pages
162
Degree date
1987
School code
0165
Source
DAI-B 48/03, Dissertation Abstracts International
Place of publication
Ann Arbor
Country of publication
United States
ISBN
979-8-207-51563-2
University/institution
University of Notre Dame
University location
United States -- Indiana
Degree
Ph.D.
Source type
Dissertation or Thesis
Language
English
Document type
Dissertation/Thesis
Dissertation/thesis number
8712954
ProQuest document ID
303617541
Copyright
Database copyright ProQuest LLC; ProQuest does not claim copyright in the individual underlying works.
Document URL
https://www.proquest.com/docview/303617541