< Back to IRCAM Forum

Real time DTW (+ HHMM GF)

Dear Mubu community,

[Question]
Is mubu.dtw usable in real-time to classify a gesture (or any other input)?

From the example, it is not clear if it can be used in real time, or how to achieve it.

[Context]
I am building a soundpainting gesture recognition patch on max and so far had good results with DTW using wekinator. I was reading a bit about the HHMM implementation of gesture follower and wanted to test its performance with respect to the wekinator DTW, but also found out about MuBu DTW that I would also like to compare (and perhaps also use a consensus algorithm between them?).

I would really like to see examples of real time use of the HHMM and DTW in mubu!

Thanks for helping,

Arthur