From: milind.a.joshi@gmail.com   
      
   Hi Jason,   
      
   What you're looking for is a good way to classify patterns. One of the   
   simplest classification routines is Naive Bayes, where each bit or each   
   row or each column of your normalized image becomes a feature, and you   
   train your classifier on a large number of such patterns.   
      
   There are many other techniques of course, like variants of the Naive   
   Bayes algorithm, a technique known as Support Vector Machines and   
   variants is the current state of the art in this type of classification   
   task. How you create your features is important. Also, note that   
   there's a difference between" handwriting" and "handprinting". The   
   difference is that with handwriting, you would have to design   
   techniques to separate parts of a joint image into smaller images, each   
   of which corresponds to a character. Note that this is not a trivial   
   task.   
      
   Generally, the three main tasks (not necessarily in that order) are:   
   0. Creating a data set, both for training and testing   
   1. Selection of features, how you extract them, how you normalize them,   
   and how you store them   
   2. Choice of algorithm, and its implementation   
   3. Testing and tuning   
      
      
   Regards,   
   Milind Joshi   
   IDEA TECHNOSOFT INC.   
   http://www.ideatechnosoft.com   
      
   [ 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)