16
$\begingroup$

The title asks it all. I have before me a data entry task that I'm not too enthusiastic about: 50-100 pages of handwritten sign out/sign in logs.

The format of the logs may help. Pages are partitioned into clearly delineated rows and columns (13r x 6c with an extra typed header row). Further helping me out is that three columns are date/time related (date, time out, time in). Also, the data in two of the columns (resource and name) are more or less enumerated, so that, for example, the name "Smith" might appear time and again in the name column, each time with the same handwriting. The last column, "Notes", is free form, but if I could automate the previous 6 columns, I wouldn't mind entering Notes by hand.

Any suggestions? (Besides 'start typing'.)

PS If there's a better SE site to ask this, let me know, I'll ask there.

$\endgroup$
  • 1
    $\begingroup$ Can you post a sample scan of the log? $\endgroup$ – Martin Thompson Mar 12 '12 at 14:17
  • 1
    $\begingroup$ Dont bother with Captricity- they ask for your email, and send u a "link email" with out a link. $\endgroup$ – user5224 Aug 14 '13 at 1:07
  • $\begingroup$ And this is what happens when you get confused with the terms «open source» and «free of charge»... that you have a company trying to get advantage on the market giving the idea that «open source» is equal to «bad». $\endgroup$ – Joan Jan 29 '18 at 18:16
6
$\begingroup$

tesseract is probbaly the best and most widespread OCR library.

It has been tested with handwriting and isn't too bad - although handwriting isn't an easy thing to read. http://arxiv.org/ftp/arxiv/papers/1003/1003.5893.pdf

$\endgroup$
  • $\begingroup$ Tesseract certainly is good option if you're looking at free/open source. It's not 100%, but it does get fairly accurate results most of the time. $\endgroup$ – Captain Kenpachi May 24 '17 at 14:12
4
$\begingroup$

If you have under 10 pages, Captricity can do it for free.

Out of the box, there are no good open source solutions to what you're looking for. Paid solutions cost a lot to license. This is based our experience building a handwriting OCR service at Captricity. We do use tesseract in production, but only as a vote that is combined with human intelligence (crowdsourcing) to deliver a high level of quality.

Hope that helps!

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.