Date: Fri, 22 Jun 2001 08:13:28 -0400

Author: "John L. Hubisz"

Subject: RE: Sutton book - Reply

Post:

Folks,

I use OmniPage 10 and rarely have errors in character recognition and this
is not the latest version! When I do get errors they are related to weird
fonts, fancy headlines, poor copies, stray marks, and so on. Sutton's book
would not present these problem as far as text is concerned.

John


>Ladies and Gentlemen,
>
>I have a couple of thoughts regarding the Sutton project.
>
>If you are going to use this book as a reference, it would be nice to be
>able to search for key words or phrases. This means that it must either
>be typed or scanned with character recognition software. With the
>state-of-the-art of character recognition software the last time I looked,
>I would not even consider it.
>
>On the other hand, you could scan the demonstration pages and type in the
>index so that you could quickly find key words, and do some sort of quick
>reference to the scanned pages so that you would not have to thumb through
>lots of scanned figures.
>
>I think you need to be careful if you scan so that you have a reasonable
>size display on the computer monitor and it comes out at a reasonable size
>when you print it. We have done this for homework solutions, and found
>the pixel rate to be about 90 dots per inch. Unfortunately, this low rate
>results in a rather crude and unprofessional looking document. If you do
>use greater pixel rates you can get nicer looking documents, but they take
>much more space and are much slower to load. The document size may be
>mitigated by going to black-and-white, but we have had more success iin
>getting a good product by using grayscale. If the original book is rather
>old, or it is a photocopy, black and white may give poor results.
>
>Overall, I would suggest investing the time and money in re-typing the
>whole thing. If you hire a good typist, it should be possible in only a
>couple of minutes per page, or less than 3 days effort. (2 minutes per
>page times 500 pages with 8 hours per day and throw in some resting
>time for the old Carpal-Tunnel) Some of the PIRA people can scan the
>pictures, make adjustments as necessary, and compose the final book.
>
>I guess another clean option is a PDF file, but it would also have to be
>re-typed. Can you search for key words in a pdf file?
>
>Perhaps the Google search engine should be applied for use with this
>file.
>
>Dick Berg

John L. Hubisz, Physics Department, Box 8202, North Carolina State
University, Raleigh NC 27695-8202; hubisz@unity.ncsu.edu, (919)515-2515,
(919)515-7331 FAX

HOME: 1604 South Salem Street, Apex NC 27502-7251, hubisz@mindspring.com,
(919)362-5782 (Voice & FAX)

Back