[Athen] OCR for Documents with Neat Handwriting?

[Shelley Haven]

Hi, Todd!

I like your explanation differentiating between the two approaches to recognition, one for existing handwritten text, the other for new text being created (where you have the additional info re: vectors, speed, etc.).  Your original question was about the former, whereas Nettie's and my responses addressed such a situation going forward (i.e., if the prof were generating new material).  And just to be complete in this discussion about handwriting recognition, let's not forget about "smart boards" (I'm using the term generically), some of which can convert neat handwriting to text either "live" or after-the-fact (e.g., mimio Studio software: <http://www.mimio.com/about_us/press_releases/2007/10_30_07.asp>) -- obviously for use with newly-generated material.

- Shelley Haven



----- Original Message -----
From: "Todd Schwanke" <tschwanke at wisc.edu>
To: "Access Technology Higher Education Network" <athen at athenpro.org>
Sent: Friday, February 20, 2009 2:35:05 PM GMT -08:00 US/Canada Pacific
Subject: Re: [Athen] OCR for Documents with Neat Handwriting?

Seems there are two types of software out there.  Ones that work with capture products (pens, tablets, etc.) and try to figure out text based on stroke information (whether real-time or after the fact) whereas others purely work on the end result (an image).  My experience, tests, searching seems to conclude so far:  

* handwriting recognition (has vector info about the stroke, including directional, pieces, speed, etc., in addition to the end result)
* OCR of handwritten text (only has a bitmap image to work from)

Of these, the handwriting recognition understandably seems to be much more accurate.  Unfortunately in this case the text is already written, so we only have scans. We probably will end up taking some alternate approaches, including audio recording.

Thanks for all the input.  Have a great weekend!
-Todd

-----Original Message-----
From: athen-bounces at athenpro.org [mailto:athen-bounces at athenpro.org] On Behalf Of Nettie Fischer
Sent: Friday, February 20, 2009 3:33 PM
To: Access Technology Higher Education Network
Subject: Re: [Athen] OCR for Documents with Neat Handwriting?

Wow, I just reread my post and didn't realize how many typos Iincluded  <ha ha> was multi-tasking
 
Another comment - The myscript software requires a pen (Pulse Scriptwriter) that costs $199.00 while the Fly Fusion pen is in the $59.00 range-a special allowed my purchase to be 39.00 including postage and handling !!! Bargain for me!.  The one thing that I see as a large difference between the two pens is the recordability option in the Pulse Scriptwriter which is not available in the Fly Fusion.
 
Nettie


On Fri, Feb 20, 2009 at 12:16 PM, Shelley Haven <rmhaven at stanford.edu> wrote:


	Along those same lines, the Windows version of Livescribe's Pulse smartpen can now do handwriting recognition with the addition of MyScript software from VisionObjects: <http://www.visionobjects.com/handwriting_recognition/pulse/pulse.htm>.
	
	- Shelley
	
	
	------------
	Shelley Haven  ATP, RET
	Assistive Technology Consultant
	Sunnyvale, CA
	408-737-2092
	


	----- Original Message -----
	From: "Nettie Fischer" <nettiet at gmail.com>
	To: "Access Technology Higher Education Network" <athen at athenpro.org>
	Sent: Friday, February 20, 2009 10:51:47 AM GMT -08:00 US/Canada Pacific
	Subject: Re: [Athen] OCR for Documents with Neat Handwriting?
	
	
	
	not software but, Fly Fusion (I purchssed a kit for 35.00 on special and ti came with 2 notebooks.  when yuo write in the notebook, it can be transcribed to print format.  Also, tablet computers offer the same option but, handwritng must be neat and of course the trick is to get proficient enough so yu dont lose downtime.
	 
	Nettie
	
	
	On Fri, Feb 20, 2009 at 10:02 AM, Todd Schwanke < tschwanke at wisc.edu > wrote:
	
	
	We have a professor who uses a hand written reader for his course that a student needs in audio format (either through recording or e-text).  It is very neatly printed (not script), but has a slight italic look to it.  Standard OCR attempts do OK on the blocky all caps titles, but when it gets to the paragraphs that look more like italic the recognition has too many errors to be edited.
	
	Before we engage on more manual attempts at converting this to e-text or a digital recording, does anyone have any software suggestions, tips/tricks, or processes that might work on fairly neat handwriting?  We've tried FineReader & Scansoft settings in Kurzweil and Microsoft Document Imaging.
	
	I can email a snippet off-list if that is helpful to see.
	
	Thanks, Todd
	
	
	_______________________________________________
	Athen mailing list
	Athen at athenpro.org
	http://athenpro.org/mailman/listinfo/athen_athenpro.org
	
	
	
	--
	Nettie T. Fischer, ATP
	Assistive Technology Practitioner
	nettiet, ATP Consultants
	www.nettietatpconsultants.com <http://www.nettietatpconsultants.com/> 
	[916] 704-1456
	
	_______________________________________________
	Athen mailing list
	Athen at athenpro.org
	http://athenpro.org/mailman/listinfo/athen_athenpro.org
	
	_______________________________________________
	Athen mailing list
	Athen at athenpro.org
	http://athenpro.org/mailman/listinfo/athen_athenpro.org
	




--
Nettie T. Fischer, ATP
Assistive Technology Practitioner
nettiet, ATP Consultants
www.nettietatpconsultants.com
[916] 704-1456


_______________________________________________
Athen mailing list
Athen at athenpro.org
http://athenpro.org/mailman/listinfo/athen_athenpro.org