Re: IFS Recess

• Previous message: A.B.P.F.: "IFS Recess"
• In reply to: A.B.P.F.: "IFS Recess"
• Next in thread: Roger L. Bagula: "Re: IFS Recess"
• Messages sorted by: [ date ] [ thread ]

Date: Sun, 08 Aug 2004 15:43:36 GMT

I, too, had noticed this. Barnsley's company seems to have disappeared
off the web!
I think it is a technology thing in which several big
companies had better presentations.....I don't see Barnsley's technique
  in the paper or the compression links.
The new jpeg standard rejected fractal compression in favor of other
methods:
( it doesn't appear fractal compression was comsidered? )
http://www.jpeg.org/
http://www.jpeg.org/jpeg2000/index.html
http://www.jpeg.org/jpeg/jpegls.html?langsel=en
http://www.hpl.hp.com/loco/
Paper: http://www.hpl.hp.com/loco/HPL-98-193R1.pdf
http://theory.cs.uni-bonn.de/~yasha/limcomplinks.html

After the success of the series of JPEG standards, the committee decided
it was appropriate to revisit the lossless coding mode within JPEG.
This mode had been a late addition to the standard, and in the baseline
form of JPEG (not using arithmetic coding) the algorithm used was not
really close to 'state of the art' techniques. In addition, it really
was not closely related to the block based DCT techniques which
characterised the lossy compression that had become the norm. Looking
at user requirements, particularly those of the medical imaging business
which was concerned about potential large errors being introduced
through lossy compression, the scope of the new standard was defined as
effective lossless and near lossless compression of continuous-tone,
grey scale and colour still images. By near lossless, it was agreed
that a scheme was needed that guaranteed a maximum error between the
original image data and the reconstructed image data.

A number of contenders for a suitable algorithm were proposed, and the
committee carried out an extensive set of measurements to determine the
best contender. The winner of these objective tests was agreed to be
the LOCO algorithm from HP Labs, whose web site offers extensive
documentation and sample code. As well as being the leading contender,
HP and Mitsubishi both agreed to offer their patents on a royalty and
license fee free basis, ensuring that implementors would hopefully not
need to make any payments to implement the new standard. As well as
offering effective compression, the algorithm was also relatively easy
to implement efficiently in PC software, and produces fast code.

A major side effect of undertaking this standards activity was that some
of the other contenders such as CALIC, FELICS, and Ricoh's CREW
algorithm in particular had some very attractive features - for example,
in the ability to provide a single codestream which could provide lossy
and lossless images without additional processing. Although outside the
direct scope of JPEG-LS, these features and the discussions they
provoked directly led to the development of the architectural approach
of the new JPEG 2000 standard.

A.B.P.F. wrote:
> Posted by request from "Fractal-World":
> http://groups.yahoo.com/group/fractal-world
>
>
>
> Date: Wed, 4 Aug 2004 12:29:56 +0400
> From: Andrey Ustyuzhanin <andrey.u@gmail.com>
> Subject: Re: IFS Recess
>
> initially I wanted to post this message to news:sci.fractals, but
> unfortunately I couldn't find appropriate USENET gate for such
> posting. May I ask you to forward my question there if it is not too
> difficult for you? There is no need to forward answers back, since I
> always can read them at groups.google.com.
> Thank you beforehand.
>
>
> austyuzhanin wrote:
> >
> > Greetings Everyone!
> > > Currently I'm studying field of IFS -based compression algorithms
> > and a bit confused by number of broken links I encounter on related
> > sites.
> > A lot of portals exist, which describe various aspects of IFS-based
> > compression, and majority of links they have lead simply nowhere
> > (Error 404, DNS name is unknown, URL is invalid, etc.) To start with
> > Michael Barnsley's iterated.com, which refuses all incoming
> > connections. Sites of other people (like
> > http://inls.ucsd.edu/~fisher/Fractals/) contain simply outdated
> > information that is rarely more recent than 90-something and a lot
> > of broken links as well.
> > > Please, tell me - is it kind of temporary disaster? Say, due to some
> > new copyright-like or government technology export restriction? Or is
> > it more like permanent state of this scientific branch? That occurred
> > due to total corruption of the IFS concept itself?
> > > --
> > Confusingly looking forward for any possible comments and answers.
> > > Andrey.
> >

-- 
Respectfully, Roger L. Bagula
tftn@earthlink.net, 11759Waterhill Road, Lakeside,Ca 92040-2905,tel: 
619-5610814 :
URL :  http://home.earthlink.net/~tftn
URL :  http://victorian.fortunecity.com/carmelita/435/

• Previous message: A.B.P.F.: "IFS Recess"
• In reply to: A.B.P.F.: "IFS Recess"
• Next in thread: Roger L. Bagula: "Re: IFS Recess"
• Messages sorted by: [ date ] [ thread ]