Path: utzoo!lsuc!watmath!thunder!water!wlrush
From: wlrush@water.waterloo.edu (Wenchantress Wench Wendall)
Newsgroups: ont.events,uw.talks,uw.cs.grad
Subject: MASTER'S ESSAY PRESENTATION
Keywords: Mr. Roland Olsson, graduate student,
Message-ID: <2499@water.waterloo.edu>
Date: 5 Jul 89 17:46:11 GMT
Distribution: ont
Organization: U of Waterloo, Ontario
Lines: 42

Dept. of Computer Science, University of Waterloo, will speak on 
``General Purpose and Binary Image Data Compression.''



DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
SEMINAR ACTIVITIES

MASTER'S ESSAY PRESENTATION SEMINAR

                    -Friday, July 7, l989

Mr.  Roland Olsson, graduate student, Dept. of Computer
Science,   University   of   Waterloo,  will  speak  on
``General Purpose and Binary Image Data Compression''

TIME:                 1:30 p.m.

ROOM:              DC 1304

ABSTRACT

Data   is  described  by  a  model  that  computes  the
probabilty  p  of  the  next  input  symbol,  which  is
optimally coded using -log2 p bits.  The predictions of
the model are used by a Huffman or an arithmetic coding
algorithm.  Techniques  based  on  the latter yield the
smallest compressed file size.

Lempel-Ziv compression algorithms are often faster, but
do  not give quite as good compression. A novel Lempel-
Ziv variant is developed.

Binary  image  compression  can  be used in for example
facsimile  machines  and  image  databases.  Run-length
coding,  READ  coding and pel colour prediction using a
template  are  reviewed. A new way of encoding an image
as a set of contours and contour locations is discussed
in  detail.  The  last method considered uses character
recognition to achieve very high compression for binary
images that mainly consist of text."