BAVI

Activities

16-22 May 2009, Santiago de Chile. First Workshop.

In the first of the three days all project members presented their respective research topics. The next two days, the discussion was centered in the different topics of the collaboration. The main activities defined during the workshop are summarized in the following.

Body and facial visual parameters extraction for animation. The parameters are obtained using the VICOM system available in the INRIA-Loria.
Study of MPEG-4 standard for body and facial animation.
Mapping of a high definition face mesh to a simple ready-to-animate face model using the MPEG4 standard. Several approaches are considered, ranging from simple manual registration techniques and to automatic ones. The main objective is the animation of arbitrary face meshes, to improve the richness and realism of the model.
Design of tracking algorithms to generate the animation parameters frame-by-frame from real video and audio.
Use of bio-inspired techniques to test symbol classification in the animated and real sequences of faces.
Multi-scale optical flow idea for motion perception was proposed.

24-29 May 2009, Santiago de Chile (Mauricio Cerda and Lucas Terissi). Research on Bio-inspired parallel architecture for motion perception. These ideas were then developed and reported in the peer-reviewed paper accepted for presentation at 4th International Conference on Bio-Inspired Models of Network, Information, and Computing Systems (BIONETICS 2009) by Cerda, Terissi and Girau.

21 September - 21 October, 2009, Nancy, France (Renato Valenzuela and Mauricio Cerda). Research on extraction of visual features from body and facial motion using the VICON system.

6-20 December 2009, Nancy, France (Lucas Terissi and Mauricio Cerda). Evaluation of markerless facial tracking algorithms using the parameters extracted by the VICON system.

27-30 October 2010, Rosario, Argentina. Second Workshop.

28 November - 5 December, 2010, Nancy, France (Lucas Terissi and Mauricio Cerda). Integration of the extracted features to the audio signal processing.

28 December 2010-28 january 2011, Nancy, France (Vanessa Peña). Stay in Inria-Loria.