[Insight-users] HAMMER registration method (Patented ?)

[Torsten Rohlfing]

First of all, thanks Luis for diligently digging up this patent
application. I have two comments.

On a general level, might it be worth for the ITK community to start
looking into gathering this sort of information, i.e., references to
patents and patent applications on algorithms of potential interest to
the toolkit? This could be done with the express intention to use the
community expertise to find prior art.

Specifically relating to HAMMER -- sorry, but none of the HAMMER
publications ever convinced me that the algorithm can do anything that
other nonrigid registration methods can't. In addition, the original
HAMMER requires an image segmentation (tissues), and segmentation errors
will necessarily generate registration errors. There is an extension to
the algorithm that I vaguely recall which works on intensity data, but
again, I can't remember seeing any evidence in any publication that this
method is superior in any way to the B-spline or Demons algorithms. I'd
gladly stand corrected if I am missing something here, but this my
current impression.

Bottom line -- is it really such a big loss for ITK that we probably
can't implement HAMMER?

TR

-- 
Torsten Rohlfing, PhD          SRI International, Neuroscience Program
 Research Scientist             333 Ravenswood Ave, Menlo Park, CA 94025
  Phone: ++1 (650) 859-3379      Fax: ++1 (650) 859-2743
   torsten at synapse.sri.com        http://www.stanford.edu/~rohlfing/

     "Though this be madness, yet there is a method in't"