[Insight-users] itkConfidenceConnectedImageFilter / Patented (part of the FuzzyConnectednessFilter) or not ?

[Luis Ibanez]

Hi Flo,

It is indeed very wise to pay attention to intellectual
"property" issues when dealing with Open Source software.


By the way, the term "PROPERTY" is incorrectly used in this
context. What many people refer to as "intellectual property"
is actually a set of exclusive rights assigned to creators
of inventions and works of art. You cannot really "own" an
idea or the expression of an idea, mainly because no idea
exist by itself, instead they are based on previous ideas.

What a patent grant is the monopoly of use of an idea for
a limited time. It practice it is only useful from preventing
others from using an idea. A Patent doesn't grant the holders
the right of exploiting the idea, because, as it is usually
the case, the idea itself may require other concepts in order
for it to work, and those other concepts may be covered by
other patents.

What copyright grants is the exclusive right of preventing
others from reproducing a work of art.

What the law grants is not the *property* of an idea or the
expression of an idea, but the benefit of having the monopoly
of exploitation of the idea *for a limited time*. The fact that
patents and copyrights expire is an indication that they are
not "property" in the proper sense.

In order to be more precise we should refer to these concepts
as "Intellectual Monopolies", not as the misnomer "Intellectual
Property".


See for example:

    Free Culture
    by Lawrence Lessig
    http://www.free-culture.cc/freecontent/
    page 56.


-----

Note that you are giving a paper publication as the reference
to a patent. This is very misleading, a Journal publication
is not the proper way of referring to a patent. The content of
a paper may be pertinent to multiple patents, or it may refer
to only some of the claims of a patent.

What you should do is to search the database of the US Patent
office:

http://patft.uspto.gov/netahtml/PTO/search-bool.html

E.g. for the terms "Fuzzy Connectedness". That will lead you
to a list including the following issued patents:


6,885,762 	Scale-based image filtering of magnetic resonance data

6,584,216 	Method for standardizing the MR image intensity scale

5,812,691 	Extraction of fuzzy object information in
                 multidimensional images for quantifying MS lesions
                 of the brain


 From that list, you could proceed to read the claims of every patent
to see how they refer to the technical aspect of the method that
you are interested on.


------


The ConfidenceConnectedImageFilter is not subject to any of these
Fuzzy connectedness patents, since it is based on a very basic concept
that anybody with a basic background on statistics will recognize as
prior art.

The filter is simply estimating means and variances of the pixel
population. There is no computation of shortest path in the
ConfidenceConnectedness filter.

Note that when you are considering the coverage of a patent,
"Generalization" of the concept is not the way to proceed.
In the way you are interpreting the method, you could extend
the coverage of Fuzzy connectedness to any region growing
method, including probably level sets; since you can always
come with a "membership function" that frames a region
growing method as a variant of the Fuzzy connectedness methods.

Instead of generalizing the claims of a patent backward,
what you must consider is the notion of prior art. That is,
whether the method under consideration is based on knowledge
that was public before the particular patent was issued.


The decadence of the patent system, and the frenzy for
appropriation of knowledge in which universities and companies
alike are obsessing, are leaving many research fields to a sad
state.

Those who care about the survival of any knowledge domain
should consider following the example of IBM in adopting
an open peer-review system for patents:

       http://dotank.nyls.edu/communitypatent/

Where patent applications are publicly reviewed online
and exposed to the scrutiny of domain experts. This novel
system will help prevent the abuses and distortions that
are becoming the norm of the patent system.



The "POCD" Patent Obsessive Compulsive Disorder is a
recent epidemic disease that plagues both industry and
academia. As more professionals get infected with it
the public domain will get depleted to the point where
you will not be able to make an addition without having
to get a license from the holder of an intellectual
monopoly.




    Regards,


        Luis



-----------
Flo wrote:
> Dear all,
> 
> As part of my PhD I'm writting a software and partially using ITK/VTK  
> functions. Although Copyright and Law infringement are not my  
> specialties, I must be careful when writing an open software for the  
> university I'm in (Sherbrooke, Canada) so that I can tell my  Supervisor 
> he can do what he wants with the code.
> 
> Thus, Does itkConfidenceConnectedImageFilter fall under any Patent,  
> especially that of Punam SAHA and UDUPA in which he detailled the  
> FuzzyConnectedFilter: Relative Fuzzy Connectedness and Object  
> Definition: Theory, Algorithms, and Applications in Image  
> Segmentation”, Nov. 2002. IEEE Transactions on Pattern Analysis and  
> Machine Intelligence. vol. 24, No. 11. p. 1485-1500.
> 
> To me, itkConfidenceConnectedImageFilter derives from the concept of   
> Fuzzy Connectedness, as there is only one step from confidence to  fuzzy 
> ... namely the Membership Function and the determination of the  
> "shortest weighted path" amongst every element to the seed point. Is  it 
> this Path Search that actually separate the Patent from any other  
> confidence Filter ?
> 
> regards,
> 
> Flo._______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>