Text Analysis

O'Gorman classifies in this section

• Skew: This is the determination of the ``slant'' of the page as scanned. For typewritten or typeset pages it is usually possible and beneficial to detect (and sometimes) remove the skew by sliding bits. In some versions of skew removal, only text-lines are straightened: individual characters remain unmodified. Our implementation deskews whole pages, and occasionally introduces a slightly disconcerting single-bit shift in the middle of a character.
• Page layout
• Character Recognition

* Graphics analysis (e.g. engineering drawing)

SECTIONS OUT OF ORDER