Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!cbnewsh!wcs
From: wcs@cbnewsh.att.com (Bill Stewart 908-949-0705 erebus.att.com!wcs)
Newsgroups: comp.compression
Subject: Re: looking for info on image compression
Message-ID: <1991Mar26.071543.18336@cbnewsh.att.com>
Date: 26 Mar 91 07:15:43 GMT
References: <5040@ns-mx.uiowa.edu> <5043@ns-mx.uiowa.edu> <12481@pt.cs.cmu.edu>
Organization: Sorcerer's Apprentice Cleaning Services
Lines: 24

]In article <5043@ns-mx.uiowa.edu> jones@pyrite.cs.uiowa.edu (Douglas W. Jones,201H MLH,3193350740,3193382879) writes:
]>...and strings of
]>pixels, reading left to right, are lousy predictors of the next pixel
]>compared to 2-dimensional neighborhoods.

In article <12481@pt.cs.cmu.edu> cokasaki@PROOF.ERGO.CS.CMU.EDU (Chris Okasaki) 
suggests using space-filling curves, and then 1-d-compressing along them.

Another cheap method, which can often provide good results
(at least for black&white), is to first take the difference
between each row and the previous one, and then use the 1-d-left-to-right
Lempel-Ziv or Huffman or whatever.  It's been a while since I did
any benchmarks, but I got decent results for gray-scale weather images;
we didn't bother using 2-D methods in our production code,
since even straight compress shrunk most of my data to 35% of original,
and color radar data compressed down to 2-3% (it was usually mostly black.)
This let us ship data at 9600 baud faster than it was arriving from
the satellite, and the application was short-term; your mileage may vary,
and better techniques could have improved the economics if we need.
-- 
# Bill Stewart 908-949-0705 erebus.att.com!wcs AT&T Bell Labs 4M-312 Holmdel NJ
(Little Girl:) When I grow up, I want to be a nurse      } From this week's UFT
(Little Boy:)  When I grow up, I want to be an engineer  } radio commercial
.... guess the Political Correctness Police don't run NYC's teachers' union yet