[<<] [<] Page 1 of 1 [>] [>>] | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subject:
Handwriting recognition on handhelds.
From: Simon Wood ####@####.#### Date: 26 Jun 2000 08:33:35 -0000 Message-Id: <44632C76B97BD211AF6B00805FADCAB202D7367E@exchange.saltaire.pace.co.uk> I've been playing with writing a handwriting recognition program for Nano-x and have had some success. The code is far from complete but does function to the extent of recording strokes and doing some simple translation of the 'point' sequence. When finished the code will released GPL (it's just too messy at present to put on general release). This work is based on libstroke, scaling input character to 5 by 5 grid and then outputs the sequence of 'cells'. grid is: A B C D E F G H I J K L M N O P Q R S T U V W X Y so AFKPU would represent a 'i' (or 'I' or '1' or '!' depending on mode) I thought that this system would work best for a 'simplified' alphabet or 'graffiti' style characters. Any comments?? I'm now at the stage of recording the whole alphabet and I'm asking for volunteers to offer time inputting characters in their own handwriting so that I can get a range of styles..... email me directly if you're interested. Simon Wood | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subject:
Re: Handwriting recognition on handhelds.
From: "Greg Haerr" ####@####.#### Date: 26 Jun 2000 15:34:09 -0000 Message-Id: <04bf01bfdf84$068dd9e0$15320cd0@gregh> : The code is far from complete but does function to the extent of recording : strokes and doing some simple translation of the 'point' sequence. When : finished the code will released GPL (it's just too messy at present to put : on general release). Simon - this is something we definitely need. However, it seems that recognition isn't an easy process. Is there any other open technology we might be able to start from, so that you don't have to reinvent all the recognition algorithms? Regards, Greg | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subject:
Re: Handwriting recognition on handhelds.
From: "Bradley D. LaRonde" ####@####.#### Date: 26 Jun 2000 15:51:26 -0000 Message-Id: <028401bfdf86$bea158c0$0701010a@ltc.com> ----- Original Message ----- From: "Greg Haerr" ####@####.#### To: "Simon Wood" ####@####.#### ####@####.#### ####@####.#### Sent: Monday, June 26, 2000 11:34 AM Subject: Re: Handwriting recognition on handhelds. > : The code is far from complete but does function to the extent of recording > : strokes and doing some simple translation of the 'point' sequence. When > : finished the code will released GPL (it's just too messy at present to put > : on general release). > > Simon - this is something we definitely need. However, it seems > that recognition isn't an easy process. Is there any other open > technology we might be able to start from, so that you don't > have to reinvent all the recognition algorithms? libstroke, Squeak, and Itsy scribble are the only ones I know about. Regards, Brad | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subject:
Re: Handwriting recognition on handhelds.
From: ####@####.#### (Justin Mason) Date: 26 Jun 2000 16:04:01 -0000 Message-Id: <200006261604.RAA10979@thalassa.netnoteinc.com> Brad said: >Greg said: >> Simon - this is something we definitely need. However, it seems >> that recognition isn't an easy process. Is there any other open >> technology we might be able to start from, so that you don't >> have to reinvent all the recognition algorithms? > >libstroke, Squeak, and Itsy scribble are the only ones I know about. Here's another -- just entering beta though. AFAICS the scribble recognition is statically built-in. --j. >From: Mike Bennett ####@####.#### >Date: Fri, 23 Jun 2000 11:42:05 +0100 >To: ####@####.#### >Subject: [ILUG] wayV: Gesture recognition as part of your desktop Righto, for those of you who may be interested version 0.0.9 (first public release) of wayV is now available at: http://wayv.sourceforge.net in the download section: http://wayv.sourceforge.net/downloads.shtml wayV is a free software project, which extends the functionality of the mouse by allowing gestures to represent actions, i.e. draw a C and xcalc starts, draw an N and Netscape starts, etc. It should be window manager independent and has been tested with Gnome, KDE and WindowMaker. Feedback would be appreciated :) Mike | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subject:
Re: Handwriting recognition on handhelds.
From: Alessandro Rubini ####@####.#### Date: 7 Jul 2000 08:11:03 -0000 Message-Id: <20000707101433.A2159@morgana.systemy.it> Hello. On Jun 26 we had a few messages about the subject matter. I have a few questions about this issue, as a few months ago I wanted to play with the problem. Can someone please give some pointer to current implementations? Are they general algorithms or complete programs for microwindows or other environment? What about the patent of Xerox? Back then when I was pondering on stroke recognition, I heard that Xerox sued 3Com for the pilot. The patent is "US4034343: Optical character recognition system". Is this an issue or not? | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subject:
Re: Handwriting recognition on handhelds.
From: Gaillard Pierre-Olivier ####@####.#### Date: 7 Jul 2000 18:45:40 -0000 Message-Id: <396624E7.C724AA6C@free.fr> Alessandro Rubini wrote: > > Can someone please give some pointer to current implementations? Are > they general algorithms or complete programs for microwindows or other > environment? - Squeak (Smalltalk) by Alan Kay seems to be very interesting because this guy knows algorithms from the 70s that CANNOT BE PATENTED (since they were published before the US patent office started delivering silly software patents). - A C'T article presented some algorithm but I don't know what it is worth and lost the reference though you may find it back by searching slashdot. - I modified a Japanese characters recognition program called kanjipad to run under Nano X (and describe a roman character-set for it). Unfortunately that program is derived from the Xerox Parc article about Unistroke. So that it offers almost no protection against a patent suit (we would win in the end, but what you are looking for is an algorithm so old that the lawyers will leave you alone). > > What about the patent of Xerox? Back then when I was pondering on > stroke recognition, I heard that Xerox sued 3Com for the pilot. > The patent is "US4034343: Optical character recognition system". > Is this an issue or not? > This is considered as an issue by Alan Kay himself. Though, he is convinced that there is plenty of prior art from the 1960s so that the patent is void. The problem is : Who is going to fight Xerox about it ? Even 3Com might prefer to pay a license... And then Xerox can still use its bogus patent against free software recognition programs (especially those that use characters made of only one stroke like graffiti)... So, the issue is pretty difficult. I would go for Squeak myself. In fact when I buy a handheld computer this autumn (probably the Ipaq from Compaq) I will evaluate Smalltalk on it, use squeak or recode Squeak. As you can see I am not even considering using the software (kanjipad) that I already have adapted : that is because you cannot make free software in the shadow of a patent. Regards, P.O. Gaillard | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
[<<] [<] Page 1 of 1 [>] [>>] |