Re: [Lcms-user] Kodak patent infringement?

[<ma...@li...>]

Hi,

Well, I'm not a lawyer, and of course this *could* have legal implications,  so, this is my tech-only point of view:


 4,500,919
========
"A system for reproducing a color original in a medium using a selected multiplicity of reproduction colorants, comprising: a
scanner for producing from said color original a train of appearance signals dependent on at least three color values of said
original; aesthetic correction circuitry for introducing aesthetically desired alteration into said appearance signals to produce
modified appearance signals; and colorant selection mechanism for receiving said modified appearance signals and for selecting
corresponding reproduction signals representing values of said reproduction colorants to produce in said medium a color match of
said modified appearance signals"

This could refer to profiles.... lcms does only implement the mechanics for profile interpretation
as described by International color consortium (ICC) www.color.org . lcms does NOT characterize
any device, and does NOT introduce any aesthetic correction. Is up to the profile (and then up
the profile generator) to embed such corrections in profile. Also, seems unapplicable to lcms, since
there is no scanner involved. Another question is lcms could be used to build such system, but lcms
does'n give this functionality out of the box.

5,432,906
=======
"A method and apparatus for generating a modified electronic image combines multiple image transforms into a single composite
transform, and then modifies the electronic image in accordance with the composite transform. The composite transform includes
sample values sequentially modified in accordance with each of the multiple image transforms being combined. The method and
apparatus further provides a user interface allowing a user to select individual transforms for composition. The use of a
composite transforms allows a user to interact with the image processing system more quickly and interactively while obtaining a
desired modified image. "

This patent is holded by Kodak, and is a bit more complex. Seems they have patented just the icc
way to handle color management... However, again, lcms does NOT implement all they describe.
lcms does NOT have any user interface (it acts as a daemon). Also, (and this is surprising, but true)
the profiles DOES NOT hold sample values, but the complete algorithm to convert between
colorspaces. lcms DOES smelt multiple image transforms into a single composite transform, but as an
option selectable by programmer. Anyway, without this smelting, color management is so slow that is
unuseable. These composite transforms are described by ICC as "device link profiles", and thus,
supported by lcms as being part of the ICC standard.


Again, don't take too seriously my comments. I'm not a lawyer, and this, of course, could cause
problems with Kodak. lcms is intended to be a portable CMM engine to be used by programmers,
not as a color management system for final users. I assume people using  it are aware lcms can be
used to build systems that breaks patents, but lcms per se isn't. Or at least that is what I belive.

Humm... If I remember correctly, java does have some similar to lcms in awt, and  seems nobody
worries about patents on that.

Regards,
Marti.



----- Original Message -----
From: "David Kaelbling" <dr...@sg...>
To: "Martí Maria" <ma...@li...>
Cc: <dr...@sg...>
Sent: Tuesday, November 13, 2001 4:11 PM
Subject: Re: [Lcms-user] Kodak patent infringement?


> Sorry for the slow response -- it took a while to find the details.  All
> we really have is FUD from our local Kodak representative, who hinted
> that if we don't renew our license with them we *might* be in violation
> of US patents 4,500,919 (excluding claim 5) and 5,432,906 if we switched
> to lcms.  If you could shed any light on this we'd greatly appreciate
> it!
>
> Thanks,
> David
>
> Martí Maria wrote:
> >
> > > I'm told that Kodak has some patents on color management.  Is lcms known
> > > to be free of them?
> >
> > ...
> >
> > I cannot assure the code is completly completly patent-free, but I have taken
> > this issue seriously. If you or anybody does know some portion of code
> > breaking any patent, please let me know. I will remove that part or provide
> > a workaround as fast as possible.
>
> --
> David KAELBLING <dr...@sg...>     Silicon Graphics Computer Systems
> 1 Cabot Rd, suite 250; Hudson, MA 01749     781.839.2157, fax ...2357
>
>