Overview of the XDOC output format in OCR Shop XTR


The XDOC format is a ScanSoft text output format which provides detailed
information about the text, images, and formatting in a recognized
document.


OCR Shop XTR offers three types of XDOC output, as specified in the
"out_text_format" parameter:
  
xdoc            Enhanced XDOC format
xdoclite        XDOC format with no format analysis
xdocplus        XDOC format with style sheet data

  
Several additional OCR Shop XTR parameters control confidence and
bounding box information:
  
xdoc_word_confidence        Output word confidences in XDOC
xdoc_char_confidence        Output character confidences in XDOC
xdoc_word_coords            Output word bounding boxes in XDOC
xdoc_char_coords            Output character bounding boxes in XDOC
 

In addition, OCR Shop XTR offers these options, which are not officially
supported:

xdoc_word_pixels   Use pixel coordinates for word bounding boxes in XDOC
no_header_footer   Do not label headers and footers
accept_thresh      Acceptibility threshold (number corresponds to the
                   confidence values seen in XDOC output)
quest_thresh       Questionability threshold


Please see the documents core12xdc.pdf and kdoctext.h for detailed
information about the XDOC format.  These documents provide enough
information to be able to parse the XDOC format.