Re: [CODE4LIB] OCR for handwritten pages
Han, Yan wrote: Hello, Colleagues, Does anyone know/use any OCR software working on handwritten pages? or at least think it is better than hiring a student key-in. I know these OCR software such as ABBYY, but they do not work on handwriting. Most 'handwriting recognition' systems are highly dependent on the script being used. Block capitals are relatively easy; idiosyncratic flowing, cursive script very hard. Interactive systems effectively train their users to write in styles legible to the system, which is not something that can be done with existing corpora. There are a number of commercial parties who do manual re-keying of handwritten pages in locations where labour is cheap, and these are likely to be your cheapest option for non-trivial volumes of text. cheers stuart -- Stuart Yeates http://www.nzetc.org/ New Zealand Electronic Text Centre http://researcharchive.vuw.ac.nz/ Institutional Repository
Re: [CODE4LIB] OCR for handwritten pages
I'm not sure if you could use reCAPTCHA or not. If you have a large enough user base for some other application and reCAPTCHA will let you specify the source document, it could be an option. http://recaptcha.net/ On Wed, Jan 13, 2010 at 2:50 PM, Han, Yan wrote: > Hello, Colleagues, > Does anyone know/use any OCR software working on handwritten pages? or at > least think it is better than hiring a student key-in. > I know these OCR software such as ABBYY, but they do not work on > handwriting. > > Thanks, > Yan > --- www.maf.org/rhoads www.ontherhoads.org
Re: [CODE4LIB] OCR for handwritten pages
Parascript (http://www.parascript.com/) has handwriting recognition software, but it only works reliably for things like forms, checks, and addresses where there is a lot of dictionary-like context to verify the image recognition. Generalized free text hand writing recognition is un unsolved problem At 01:50 PM 1/13/2010 -0700, Han, Yan wrote: Hello, Colleagues, Does anyone know/use any OCR software working on handwritten pages? or at least think it is better than hiring a student key-in. I know these OCR software such as ABBYY, but they do not work on handwriting. Thanks, Yan
Re: [CODE4LIB] OCR for handwritten pages
Perhaps this isn't substantially different from student key-in, but handwriting recognition may be a good task to outsource to Mechanical Turk: https://www.mturk.com/mturk/welcome Good luck, -Mike On Wed, Jan 13, 2010 at 15:50, Han, Yan wrote: > Hello, Colleagues, > Does anyone know/use any OCR software working on handwritten pages? or at > least think it is better than hiring a student key-in. > I know these OCR software such as ABBYY, but they do not work on handwriting. > > Thanks, > Yan >
Re: [CODE4LIB] OCR for handwritten pages
Does anyone know what Evernote [http://www.evernote.com/about/home.php] uses as their back-end image recognition engine? There are a lot of testimonials out there which claim that Evernote can capture the text from pictures of white-boards and napkins. I'm sure that mileage will vary but might be worth checking out. You can get at their web-services (including the text extracted from images) by applying for a developers key [http://www.evernote.com/about/developer/api/] and worse case the cost is $5/month or $45/year for premium uploads (500MB/month). Regards, Tricia Han, Yan wrote: Hello, Colleagues, Does anyone know/use any OCR software working on handwritten pages? or at least think it is better than hiring a student key-in. I know these OCR software such as ABBYY, but they do not work on handwriting. Thanks, Yan
Re: [CODE4LIB] OCR for handwritten pages
There was some work done in the UMass CS Dept[1] a long time ago. I'm not aware of any end-user software available, though some proprietary systems like Evernote[2] have pretty advanced text in image recognition capabilities. The high accuracy necessary for recognizing the text of entire documents is probably a very serious hurdle for technology like this. [1] http://orange.cs.umass.edu/irdemo/hw-demo/ [2] http://www.evernote.com/ Best, Aaron On 1/13/2010 3:50 PM, Han, Yan wrote: Hello, Colleagues, Does anyone know/use any OCR software working on handwritten pages? or at least think it is better than hiring a student key-in. I know these OCR software such as ABBYY, but they do not work on handwriting. Thanks, Yan -- Aaron Rubinstein Digital Project Manager W.E.B. Du Bois - Verizon Digitization Project Special Collections and University Archives University of Massachusetts, Amherst Tel: (413)545-9637 Email: arubi...@library.umass.edu Web: http://www.library.umass.edu/spcoll/
[CODE4LIB] OCR for handwritten pages
Hello, Colleagues, Does anyone know/use any OCR software working on handwritten pages? or at least think it is better than hiring a student key-in. I know these OCR software such as ABBYY, but they do not work on handwriting. Thanks, Yan