home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!corton!chorus!octave.chorus.fr!jloup
- From: rthomson@dsd.es.com (Rich Thomson)
- Newsgroups: comp.compression
- Subject: Fractal compression
- Message-ID: <14026@chorus.fr>
- Date: 13 Aug 92 14:16:06 GMT
- Sender: jloup@chorus.fr
- Reply-To: rthomson@dsd.es.com (Rich Thomson)
- Organization: Design Systems Division, Evans & Sutherland, SLC, UT
- Lines: 101
-
- [Rich Thomson <rthomson@dsd.es.com> sent me the following message
- in answer to item 17 of the FAQ "What is the state of fractal compression?".
- I thought it was interesting and Rich asked me to post it.
- Jean-loup Gailly <jloup@chorus.fr> ]
-
-
- At SIGGRAPH 1992, Barnsley was at a press conference held by the
- publisher Jones & Bartlett. They are publishing two new books by
- Barnsley that give more concrete details of the method:
-
- The Fractal Transform,
- Michael F. Barnsley and Louisa F. Anson
- ISBN 0-86720-218-1, ca. 250 pp, $49.95
-
- and
- Fractal Image Compression
- Michael F. Barnsley and Lyman P. Hurd
- ISBN 0-86720-457-5, ca. 250 pp., $49.95
-
- the publisher can be reached at:
-
- Jones & Bartlett Publishers, Inc.
- One Exeter Plaza
- Boston, MA 02116
- (617) 859-3900
- FAX: (617) 859-7675
- kpeters@math.harvard.edu
-
- Also, at this press conference Barnsley claimed to have patents filed
- "worldwide" for their technology.
-
- >from Tal Kubo <kubo@zariski.harvard.edu>:
- >
- >According to Barnsley's book 'Fractals Everywhere', this method is
- >based on a measure of deviation between a given image and its
- >approximation by an IFS code. The Collage Theorem states that there is
- >a convergent process to minimize this deviation. Unfortunately,
- >according to an article Barnsley wrote for BYTE a few years ago, this
- >convergence was rather slow, about 100 hours on a Cray, unless assisted by
- >a person.
-
- Their performance has improved dramatically beyond what they were
- talking about in BYTE a few years ago. Human assistance to the
- compression is no longer needed and the compression time is
- reasonable, although the more time and compute power you throw at the
- compression, the smaller the resulting file for the same level of
- quality.
-
- >Barnsley et al are not divulging any technical information beyond the
- >meager bit in 'Fractals Everywhere'.
-
- See above; now that they have obtained their patents they are talking
- more about the method (since their 'competitive edge' is protected).
-
- >There is reason to believe that Barnsley's company has
- >*no algorithm* which takes a given reasonable image and achieves
- >the compression ratios initially claimed for their fractal methods.
-
- Barnsley may have pissed alot of people off because of the way he
- presents (or doesn't) his work, but that doesn't mean that his method
- is bogus. It is quite reasonable and it works.
-
- >Steve Tate <srt@duke.cs.duke.edu> confirms:
- >
- >Compression ratios (unzoomed) seem to range from 20:1 to 60:1... The
- >quality is considerably worse than wavelets or JPEG on most of the
- >non-contrived images I have seen.
-
- You need to supply a date for this observation because Iterated
- Systems Inc. has 10 full-time mathematicians who do nothing but work
- on making the fractal compression algorithms more efficient as well as
- produce higher-quality images for a given file size. I suspect that this
- observation "quality is considerably worse" is quite dated by now.
-
- > A. Jacquin, 'Fractal image coding based on a theory of iterated
- > contractive image transformations', Visual Comm. and Image
- > Processing, vol SPIE-1360, 1990.
-
- For people who are starving to find out what the method really is,
- this is the best paper that explains the concept in a simple way.
-
- The algorithm searches for 'affine redundancy' by searching for
- transformations that map regions of pixels to a smaller region of
- pixels in the source image. The space of all possible transformations
- is quite large for any given image, so that is why the process needed
- human aid at first and also explains why smaller file sizes can be
- found for a given quality level if more compute power/time is applied
- to the search process.
-
- Also, at the press conference Barnsley stated that Microsoft has
- licensed the technology and will be supplying a CDROM with 16,000
- compressed images on it. He didn't mention what the original
- resolution of the images was, but I suspect something common for PC
- images like 640x480.
-
- -- Rich
- --
- Repeal the personal income tax; vote Libertarian in 1992.
- Disclaimer: I speak for myself, except as noted; Copyright 1992 Rich Thomson
- UUCP: ...!uunet!dsd.es.com!rthomson Rich Thomson
- Internet: rthomson@dsd.es.com IRC: _Rich_ PEXt Programmer
-