From: christian.stieg@gmail.com   
      
   Hi everyone,   
      
   I hope I'm not all the way wrong here with my problem.   
   I am researching face recognition here at the university of ilmenau,   
   and I want/have to do it with hidden-markov-models.   
      
   I implemented the HMM-code using mainly the rabiner-tutorial,   
   and it seems to work. (discrete HMM work definitely, continuous   
   I am not THAT sure, does anyone know a way to "test" this,   
   to verify that the algorithm is correct?)   
      
   When I do the training with my observation, created with   
   wavelet decomposition of picture blocks, after some iterations   
   of the EM-algorithm I get an error because the values of the   
   gaussian distributions go zero. I calculate these values with:   
      
   dens= 1/sqrt((2.*pi).^D.*norm(Sigma)) * exp((-0.5)*vec'*inv(Sigma)*vec)   
      
   vec being O-mu.   
   I found out that they go zero, because the exp-term is negative,   
   and e^ a negative number is zero, so the whole thing is zero.   
      
   The problem is, I don't know why that happens.   
   Can anyone give me a hint, what could be wrong or what I could do?   
   If you need code, just tell me, I'll show you. :)   
      
   Kox   
      
   [ 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)