Accelerometer-based gesture recognition

Question

I'm developing an embedded accelerometer-based hand gesture recognition. Does anybody know about some free libraries to employ or to start from? I'm working with embedded linux and I'm looking for something not too much computation-heavy. I couldn't find much on google apart from dozens of theoretical papers, youtube videos or "closed" applications.

I understand it's a pretty broad question, but I think it can help everybody in the same situation.

Thank you!

Answer 1

If you haven't already seen these...

There are two projects started as academic projects:

http://www.dfki.de/~rnessel/tools/gesture_recognition/gesture_recognition.htm (with source code)

and

the gRmobile framework: all I can find is a paper: http://www.ic.uff.br/~esteban/files/papers/SBGames09_Mark_A.pdf

Hope that helps! There doesn't seem to be much out there. But, if you are looking to do a little more work, not necessarily relying on a library, definitely be sure to read through the gesture information from the Android Dev site (the Gesture Recognition API): http://developer.android.com/resources/articles/gestures.html

Answer 2

This question is already answered, but for the sake of posterity, here goes:

http://www.wiigee.org/

Meant for use with the Wiimote, wiigee is an open-source Java based implementation for pattern matching based on accelerometer readings. It accomplishes this using Hidden Markov Models[1].

It was apparently used to great effect by a company, Thorn Technologies, and they've mentioned their experience here : http://www.thorntech.com/2013/07/mobile-device-3d-accelerometer-based-gesture-recognition/

Alternatively, you could consider FastDTW (https://code.google.com/p/fastdtw/). It's less accurate than regular DTW[2], but also computationally less expensive, which is a big deal when it comes to embedded systems or mobile devices.

[1] https://en.wikipedia.org/wiki/Hidden_Markov_model
[2] https://en.wikipedia.org/wiki/Dynamic_time_warping