Re: Searching for circular structures in an image

*To*: mathgroup at smc.vnet.net*Subject*: [mg130606] Re: Searching for circular structures in an image*From*: debguy <johnandsara2 at cox.net>*Date*: Thu, 25 Apr 2013 02:55:24 -0400 (EDT)*Delivered-to*: l-mathgroup@mail-archive0.wolfram.com*Delivered-to*: l-mathgroup@wolfram.com*Delivered-to*: mathgroup-outx@smc.vnet.net*Delivered-to*: mathgroup-newsendx@smc.vnet.net*References*: <kl8e4c$on3$1@smc.vnet.net>

Good question i'm not sure I know. I looked on wikipedia to decipher what your asking. There is a recent posting just days ago involving shape recognition. However the "answer" is also not mathematical theory as, mathematically, many kinds of errors arise in the math is very assuming (ie, drops many sol'n). Yet it does get one in the ball-park easily and quickly one can say. What I found on wiki didn't qualify as a function (none was posted, and it wouldn't be possible to post one. Nor did the problem or answer wiki describes qualify as formal mathematical theory or anything close to being a theory. Infact it only describes a whimsical comparison and doesn't do justice to what kind of math might be needed. A "generalized" sol'n might be to find linear multiple of splines in 3 space within a given percentage, not just ellipsoids in 2 space? There are many recognition and learning algorithms. Which brings up an important question: what exactly do the data and results require? It's a popular topic and Im sure google is bound to have many results if your looking for code and algorithms. As far as pattern recognition i'm unsure what if any of mathematica's new features best fit. It seems to be rather straight forward, statistically speaking, given the previous posting. I'm sure mathematica will prove useful while doing it.