Print

Print


SignWriting List
January 27, 2011

Hello Everyone -
Below is more information about the Automatic Hand Gesture and Movement Recognition software development by Gan Lu. Gan gave me permission to post this information last January as you can read below, and I never got to posting it. Very sorry for the delay, but here it is... Thank you again, Gan. I hope someday it will become finished software by others... Val ;-)

--------

On Jan 27, 2011, at 6:54 AM, Gan Lu wrote:

> Hello Valerie;
> 
>    Yes, sure! No problem for me. Thanks for your interest and your sharing information.
> 
>   Gan :-)
> 
> >>> Valerie Sutton <[log in to unmask]> 27/01/11 2:47 PM >>>
> Hello Gan and Geunhee!
> Thank you for this discussion - I only wish the discussion could be on the SignWriting List, because I know of at least one other programmer who might be working on this issue and it would be informative for others too...Do I have your permission to post this to the SignWriting List, or excerpt it, and then I can answer? Thank you for your help with this - I am continually looking for ways to inform SignWriting List members of projects around the world - Val ;-)
> 
> ------
> 
> 
> On Jan 27, 2011, at 5:03 AM, Gan Lu wrote:
> 
> > Hello all;
> > 
> > Thanks Val for forwarding this email to me.
> > 
> > Hi, GeunHee. Concerning about you question of the input methods for SignWriting, based on our knowledge, we classified them into 
> > 
> > A. Human-force input method : Hand writing or drawing (Time consuming)
> > B. 
> > 1. Indirect Sign Writing input methods (using keyboard or mouse control: Time consuming and need to transcribe the gestures into SW symbols in mind first, which is not direct) 
> > 2. Direct Sign Writing input methods (Directly using the hand gestures or facial expression as input. By automatically recognise them using some HCI systems, the corresponding SW symbols can be obtained for further use, i.e., display or information disseminating. )
> > 
> > Obviously, the Direct SW input method has advantages compare to other input methods. Currently, we are implementing a direct Sign Writing system based on the hand gesture and movements. 
> > 
> > There are two approaches can be considered to achieve the Direct Sign Writing, vision based (photo the hand gestures by camera) and data glove based (the hand gesture data is obtained by the sensors attached on the hand). Compare to the latter, the fomer has advantages in terms of less-clumsy. But it has some inherent difficulties such as occlusion problem which is difficult to overcome by current techniques. So we adopted the glove based approach. 
> > 
> > Therefore, our research is to use a pair of data gloves to gather the hand gesture and movement data, recognise them by the developed algorithms, and finally display the corresponding Sign Writing symbols on to a large screen. To our best knowledge, this is the first ever designed system for Direct Sign Writing. 
> > 
> > The developed system is now able to recognise different hand gestures, movements on wall/floor planes, movements on 8 directions, linear/non linear movements, and repeating movements. We developed a visualisation interface to visualise the recognition output. As we may understand, the recognition of movement is very difficult due to since the movement can exist in arbitrary form and shape, so this is going to be a long term project for us to fully implement the hand movement recognition for Direct Sign Writing. Also, our university is also doing some working in the facial expression recognition. Thus, in the future, we may integrate both the hand and face together for Direct Sign Writing. 
> > 
> > 
> > However, as I said, our project is a long-term ongoing research project, and it may not mature enough to be considered for industrial product so far (which is what your interest). 
> > 
> > 
> > Now, if we come back to the indirect sign writing methods, we used to use the keyboard and mouse as the input methods. However, as you mentioned the smart phones. I may think there should be a way to design a smart phone built-in hand writing system for SW. Therefore, by a user uses his/her finger to draw a SW symbol on the touch phone screen (like people interacting with their iphone), the built-in recognition system should be able to recognise the users hand drawing and display the corresponding SW symbols. Is that what you looking for, Geunhee? I don't think anyone has done anything on that though....
> > 
> > I don't know if I answered your question?
> > 
> > 
> > Gan
> > 
> > 
> > 
> > 
> > >>> Valerie Sutton <[log in to unmask]> 27/01/11 4:04 AM >>>
> > Hello GeunHee Lee!
> > Thank you for this email. I am sharing it with Gan Lu, from the University of Central Lancashire in the UK, who has developed software that sounds a little bit like what you are suggesting in your email below....
> > 
> > Gan Luís doctoral work is in software development, working on a project called Automatic Hand Gesture Recognition for SignWriting, which uses a glove to produce sign language movement written in SignWriting.
> > 
> > Gan is at the University of Central Lancashire in the UK...
> > 
> > Gan, is your work similar to what GeunHee is talking about?
> > 
> > Val ;-)
>