andrel wrote:
> Ok, we agree that you can not average them as bitmap. So you need a 
> model. Which means extracting parameters from the fingerprint in a way 
> that they can be averaged. With vector I assume that you think that you 
> can normalize all fingerprints and then at some sampling points compute 
> the angle of the closest line in the fingerprint or something like that. 
> Then you can for each point define the average direction of 
> fingerprints*. Only, what you get is the average model not an average of 
> the input. Don't confuse a model with reality.**

Sir Francis Galton did just that; came up with a string of letters and 
numbers to represent a fingerprint, with a 1:10,000,000,000,000 
collision rate.  Plenty for humans.

-- 
Tim Cook
http://empyrean.digitalartsuk.com

-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GFA dpu- s: a?-- C++(++++) U P? L E--- W++(+++)>$
N++ o? K- w(+) O? M-(--) V? PS+(+++) PE(--) Y(--)
PGP-(--) t* 5++>+++++ X+ R* tv+ b++(+++) DI
D++(---) G(++) e*>++ h+ !r--- !y--
------END GEEK CODE BLOCK------

Post a reply to this message