alexgold87: iOS (ipad/iphone/ipod) doesn't support WebGL *at all* for the moment. The only WebGL piece you can do with HTML/JS for the moment is an iAd...

chrisbrobs : I'm a bit behind my schedule, but I didn't forgot you ! ;)

If you can avoid inputting duplicate point, it's going to speed the cleaning of the shape byt the lib (but I don't how much milliseconds are gained, really).

When you add a new variant, you don't need to clear, because clear empty the list from all user input. I need to add that on the original post.

chrisbrobs : strange, that means that between the moment I was editing the various files and the moment I uploaded, I forgot to save the implementation of one of the actions.

I'm going to check that this afternoon and reupload it...

newt : I saw on IRC you proposed an idea to compare old/new points, but I'm not sure where you want to add that.

If it's an idea to simplify new gestures before adding them to the library, it's not necessary. You can use a bunch of points (even if some of them are overlapping), the lib is doing itself the cleaning (it's a kind of Procruste analysis to simplify the input gesture to compare it against the reference).

Some screenshots from the paper :

1: you can see there's some variability in the input gesture

<img src="http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/01.png" border="0" />

2: so the gesture is resampled internally to a predefined number of equidistant points

<img src="http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/02.png" border="0" />

3: the gesture is rotated, always in the same direction, to make it

rotation independent

<img src="http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/03.png" border="0" />

4: finally some error rate regarding recognition, depending on the number of availble samples to compare against (you see there that 3 is the minimum efficient number) - Rubine is another algorithm from the field, and DTW is Dynamic Time Warping, more powerful, but way more ressources hungry (too much for a quick Javascript, at least)

<img src="http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/04.png" border="0" />

2Bdigital : I'm not sure, but does the Unistroke Gesture Recognition plugin helps you in that context (by adding a swipe gesture) ?

newt : I'm not really sure what you mean by "the default gestures cause a false positive sometimes. Is there any way to take the defaults out?"

If, when you draw something that's not looking something you aleardy have in the library of gesture, the lib will try to match it anyway (but it will get a very bad score). So you can filter false positives with a high score. Is it helping you ?

All : I didn't think about that in the beginning, but with that lib, you can do a "swipe gesture" detection on mobile ;).

I also edited the original post to add something : to have a good recognition when you add new gestures, you need to add at least 3 variants of that gesture.

Since JesseJoh is busy with his work, I'm placing here an addition I added to his plugins : the possibility to rotate the 3 elements (<div>, <iframe> and <img>). It doesn't work in IE,though.

I also added a property field to add an "ID" to those tags, to let you access them from outside C2's framework.

When JesseJoh is back, I hope we could integrate them in his plugins (for the moment, I added an extension "_Pode" to the modded plugins, to let everybody use the version they want, without breaking .capx project in the future).

In the end, you should thanks JesseJoh one more time for his work !

<img src="http://dl.dropbox.com/u/1412774/demo.png" border="0">

The plugin archive :

http://dl.dropbox.com/u/1412774/pode_html_pack%28from_JJ%29.zip

The demo :

http://dl.dropbox.com/u/1412774/html_pack_pode_example.capx

vidi: is it a problem related to Retina graphics (and then dividing pixel size by two in each direction) ?

I just added two things to the original post I'm just remembering now...

Edit : <font color="red">Update 08 - february - 2012</font>

New build !http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/pode_1StrokeGestureRecognizer.0.3.zip

Demo : http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/index.html

To appologize, there's a new functionnality : remove a specific gesture, by name

Edit : <font color="red">Update 07 - february - 2012</font>

New build ! http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/pode_1StrokeGestureRecognizer0.2.zip

• Centroid, Bounding Box, Angle
• delete all gesture from the lib (even those given at the start)
• add a custom list of points ! (provided as "X1,Y1|X2,Y2|...|XnYn)

The demo is updated, by the way : http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/index.html

Edit2 : I also added a Custom Gesture Creator : http://dl.dropbox.com/u/1412774/CustomGestureCreator/index.html

http://dl.dropbox.com/u/1412774/CustomGestureCreator/GestureCreator.capx. It's easier for the creation of custom gesture's pointlists...

____________________

I remembered using a library to work with wiimote, helping to recognize gestures made with the gamepad.

That lib is there : http://depts.washington.edu/aimgroup/proj/dollar/

So I made a plugin for C2 today, that helps to recognize gestures made by the user, which uses that lib.

You can get that plugin there : http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/pode_1StrokeGestureRecognizer.zip

I also made a quick demo app, mimicking the one from the original link :

<img src="http://i.imgur.com/6fR9V.jpg" border="0">

http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/index.html

And a demo capx : http://dl.dropbox.com/u/1412774/oneStrokeGestureRecognizerC2Plugin/oneStrokeGestureRecognizerC2PluginDemo.capx

One thing to know : in the original algorithm, the recognition works when there's more than 10 points defined for a gesture.

In a future rev., I'm going to add the possibility to give directly to the plugin a list of points (for example, if you already determined a predefined gesture), and the possibility to load that from a file.

When I have more time, I'm going to integrate the N strokes (multistrokes) Javascript lib from the same team...

Edit : it's up to you to set a level (score) under which the gesture isn't recognized, because the lib will try its best to recognize something anyway. I usually go for something between 75 (it let pass a few bad detection) to 90 (pretty strict detection).

Edit2 : In the original paper, to have a good recognition when you add new gestures, you need to add at least 3 variants of that gesture. So if you want to detect a new gesture, to forget to draw it 3 times and add it 3 times ! :)

Edit3 : the action "Clear Point List" empties any user defined gesture, not gesture pre-packaged with the lib !