On Tue, Mar 11, 2008 at 6:19 AM, Ramakrishnan Sundaram
<[EMAIL PROTECTED]> wrote:
> On Tue, Mar 11, 2008 at 6:40 PM, Gautam John <[EMAIL PROTECTED]> wrote:
>
>  >  Ideally, we'd be able to strip out the original text and lay the
>  >  translated text in that very same area.
>
>  That can't be automated.

That's a very bold and categorical claim... it may not be practical
for Gautam to automate it, but it's certainly doable. In fact most OCR
software these days does it as a matter of course. It will identify
the text areas and image areas as rectangular regions, and allows you
to select the text independently of the images. If you convert to Word
format it preserves the text block flow, which should allow you to
replace the text independently of the graphics. I know PDF has the
same capability, and Acrobat 8.1 pro will do OCR from scanned images,
but I don't know if it creates text blocks or allows you to reflow the
text. I believe it does though.

-- Charles

Reply via email to