Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.compression > #1556 > unrolled thread

Trinary image compression

Back to article view | Back to comp.compression

Contents

  Trinary image compression glen herrmannsfeldt <gah@ugcs.caltech.edu> - 2012-10-22 19:22 +0000
    Re: Trinary image compression Willem <willem@turtle.stack.nl> - 2012-10-22 20:53 +0000
      Re: Trinary image compression glen herrmannsfeldt <gah@ugcs.caltech.edu> - 2012-10-22 23:43 +0000
        Re: Trinary image compression Niels Fröhling <spamtrap@adsignum.com> - 2012-10-22 20:41 -0600
      Re: Trinary image compression Thomas Richter <thor@math.tu-berlin.de> - 2012-10-23 08:39 +0200
        Re: Trinary image compression glen herrmannsfeldt <gah@ugcs.caltech.edu> - 2012-10-23 06:59 +0000

#1556 — Trinary image compression

I very much like the idea of JBIG and the compression of binary
(black, white, but no gray or other colors) images.

There are, however, some books, at least computer manuals,
written with two ink colors. (Black and blue for example.)

There is, as far as I know, no JTIG compression for trinary
(black, blue, white) images. (Presumably one could allow for
any three colors, likely one being white.)

More specifcally, there are many DEC manuals in bitsavers
that have lost some bits in scanning. If there was a format
to store them in, there might be more interest in doing it.

-- glen

[toc] | [next] | [standalone]

#1557

glen herrmannsfeldt wrote:
) I very much like the idea of JBIG and the compression of binary
) (black, white, but no gray or other colors) images.
)
) There are, however, some books, at least computer manuals,
) written with two ink colors. (Black and blue for example.)
)
) There is, as far as I know, no JTIG compression for trinary
) (black, blue, white) images. (Presumably one could allow for
) any three colors, likely one being white.)
)
) More specifcally, there are many DEC manuals in bitsavers
) that have lost some bits in scanning. If there was a format
) to store them in, there might be more interest in doing it.

I think it would be easiest if you split the image in two,
one containing the black text and one containing the blue text.


SaSW, Willem
-- 
Disclaimer: I am in no way responsible for any of the statements
            made in the above text. For all I know I might be
            drugged or something..
            No I'm not paranoid. You all think I'm paranoid, don't you !
#EOT

[toc] | [prev] | [next] | [standalone]

#1558

Willem <willem@turtle.stack.nl> wrote:

(snip, I wrote)
> ) I very much like the idea of JBIG and the compression of binary
> ) (black, white, but no gray or other colors) images.

> ) There are, however, some books, at least computer manuals,
> ) written with two ink colors. (Black and blue for example.)

> ) There is, as far as I know, no JTIG compression for trinary
> ) (black, blue, white) images. (Presumably one could allow for
> ) any three colors, likely one being white.)


(snip)

> I think it would be easiest if you split the image in two,
> one containing the black text and one containing the blue text.

Yes, that is fine. But there are no standard compression algorithms,
more specifically, no popular decompression algorithms, that do it.
(Considering some cases where the decompression algorithm is
specified separately from the many possible compression algorithms
that it supports.)

I suppose one could use LZW on two bit/pixel image files, but that
is likely not near as good as JBIG2 on the two separate images.

There may be use for four or five color, but no shading, in some
image systems. 

The nice thing about JBIG2 is that it has been supported by PDF
for some years now. 

-- glen

[toc] | [prev] | [next] | [standalone]

#1559

> Yes, that is fine. But there are no standard compression algorithms,
> more specifically, no popular decompression algorithms, that do it.

There are a lot of papers about palette-based compression. Most of the research 
beats JBIG by a margin in it's own territory, b/w in the end is just a two-color 
palette.
In your three color case, if you have no better inspiration, just code two 
planes (a: is it color 0 or the others?, b: is it colour 1 or the others?, c: is 
implicit).

Ciao
   Niels

[toc] | [prev] | [next] | [standalone]

#1560

On 22.10.2012 22:53, Willem wrote:
> glen herrmannsfeldt wrote:
> ) I very much like the idea of JBIG and the compression of binary
> ) (black, white, but no gray or other colors) images.
> )
> ) There are, however, some books, at least computer manuals,
> ) written with two ink colors. (Black and blue for example.)
> )
> ) There is, as far as I know, no JTIG compression for trinary
> ) (black, blue, white) images. (Presumably one could allow for
> ) any three colors, likely one being white.)
> )
> ) More specifcally, there are many DEC manuals in bitsavers
> ) that have lost some bits in scanning. If there was a format
> ) to store them in, there might be more interest in doing it.
>
> I think it would be easiest if you split the image in two,
> one containing the black text and one containing the blue text.

Just FYI, JBIG2 supports exactly that. We started an AMD on this at 
around 2006 and it should be long complete by now. I do not remember 
exactly how it performed.

So long,
	Thomas

[toc] | [prev] | [next] | [standalone]

#1561

Thomas Richter <thor@math.tu-berlin.de> wrote:

(snip, someone wrote)

>> I think it would be easiest if you split the image in two,
>> one containing the black text and one containing the blue text.

> Just FYI, JBIG2 supports exactly that. We started an AMD on this at 
> around 2006 and it should be long complete by now. I do not remember 
> exactly how it performed.

I will have to read through it again.

Then the problem is finding the right scanner to generate the scans.

-- glen

[toc] | [prev] | [standalone]

Back to top | Article view | comp.compression

csiph-web