XPost: comp.ai.neural-nets, sci.image.processing   
   From: loerge@gmx.at   
      
   JinSoo Kim wrote:   
   > Hi,...   
   >   
   > I want to develop a fast and simple rigid graph matching algorithm.   
   > Let me explain the problem first..   
   >   
   > There is a rigid graph with 40-80 nodes. The 'rigid' means that the   
   > positions   
   > of all nodes are not changed significantly. It can change its position   
   > only slightly. (In my research, such a graph is the minutiae in the   
   > fingerprint   
   > image.)   
   >   
   > The whole graph can be rotated or translated. (For example, when   
   > somebody press his fingerprint on the fingerprint image sensor, the   
   > fingerprint image may   
   > be rotated or translated)   
   >   
   > And some nodes may be lost or some spurious nodes may be added.   
   > Sometimes, two graphs which are to be compared may have small area of   
   > common   
   > subgraph.   
   >   
   > In this case, is there any known fast, simple and efficient algorithm   
   > which measures the similarity of the two graphs?   
   > Any pointing to the references will be much appreciated.   
   >   
   > Thanks in advance...   
      
   i've recently read an article about recognition of fingerprints using   
   the frequency content of turning functions of the minutiae. something in   
   citeseer (http://citeseer.nj.nec.com/cs) i guess. might be worth a try.   
      
   [ comp.ai is moderated.  To submit, just post and be patient, or if ]   
   [ that fails mail your article to , and ]   
   [ ask your news administrator to fix the problems with your system. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)