An audio-driven dancing avatar, F. Ofli, Y. Demir, Y. Yemez, E. Erzin, A.M. Tekalp, K. Balci, I. Kizoglu, L. Akarun, C. Canton- Ferrer, J. Tilmanne, and E. Bozkurt, Journal on Multimodal User Interfaces, vol. 2, no. 2, pp. 93-103, September 2008.

Audio-driven human body motion analysis and synthesis, F. Ofli, C. Canton-Ferrer, J. Tilmanne, Y. Demir, E. Bozkurt, Y. Yemez, E. Erzin, A.M. Tekalp, ICASSP’08 Las Vegas, USA.

1.    Body motion capture and tracking demo (for salsa dance):

 

2.    Body motion capture and tracking demo (for belly dance):

 

3.  First scenario – input audio contains both salsa and belly:

 

4. Second scenario a) – input audio contains slow and fast parts from the salsa audio type:

 

5.  Second scenario b)- input audio contains slow and fast parts from the belly audio type:

 

6.  Third scenario – input audio is different than both of the training audio types, i.e., neither salsa nor belly: