F1  Train New Font
       To build a new font, input the image file of text which contains the
       particular typestyle you want to train. You may either type in the 
       image filename manually or use the [TAB] key for selection. Make sure 
       the page is clean and the characters are intact. Then enter the font  
       file to be created. An automatic .FNT extension will be given.

       Note : The most efficient way to train a new font is to run 
              Interactive OCR with a partially-built font file. 
              The partial font file may contain as few as 5 to 10 
              trained characters. Use [F1] Train New Font in this Menu 
              to train maybe a line of characters, then exit font training 
              and run Interactive OCR at a high recognition sensitivity on 
              this partial font. 

F2  Expand Existing Font
       This option is not frequently used but is helpful for retraining 
       mis-recognized characters in an existing font file. To expand an 
       existing font, enter the names of the scanned image and the font file 
       that you would like to expand, again, you may input both filenames 
       either manually or select from the list with the [TAB] key.

F3  Edit font file
       Sometimes, mis-recognized characters are due to erroneous training.
       This option allows you to remove an erroneously trained character. 
       First, input the name of the font file, then enter the character to be
       removed. This is a global removal, therefore, retraining is required.