DTW for 3D gesture recognition

Question

I'm using DTW to compare gestures in 3D space, relying on 3axis accelerometer data, using Python MLPY module.

I'm in doubt whether I need to apply DTW to each axis (x, y, z) independently and then sum up the resulting costs or there is some way to combine the axis before running DTW. I think just running DTW on the norm of the vector is misleading, as this way you would just discard useful information.

What would you suggest?

score 0 · Accepted Answer · edited May 23 '17 at 12:33

0

I would use the Eucledian distance, see at the bottom of my earlier answer.

edited May 23 '17 at 12:33

Community

1
1

answered Feb 11 '12 at 17:19

Ali

56,466
29
168
265

Thank you! I think I'll do this way. I'm looking for a good C implementation of DTW, do you have any suggestion? – stef Feb 16 '12 at 08:49
@stef No, unfortunately, I could only Google too... :( – Ali Feb 16 '12 at 09:48
Same for me, I found this http://www.phon.ox.ac.uk/files/slp/Extras/dtw.html but it gives "Segmentation fault" even when I run it with the example data.. – stef Feb 17 '12 at 16:39

DTW for 3D gesture recognition

1 Answers1