Current position
I am an Assistant Professor in the M2S Lab of the University of Rennes 2 working on virtual reality used to analyse sports. These studies deal with virtual reality, biomechanical analysis and perception. I am also a Scientific Collaborator in the Bunraku team of the IRISA working on virtual humanoids animation. These works are based on real-time motion adaptations. Some of these works were embedded in the MKM animation engine. This software is now selt by the Golaem company in which I work as a Scientific Advisor.
Research interests:
- Virtual reality
- Animation of humanoids
- Biomechanics
Education
| 2005 |
PhD in Computer Sciences – Bunraku team INSA (Institut National des Sciences Appliquées), Rennes, France |
| 1996 |
Engineer in Computer Sciences (speciality Image synthesis) INSA (Institut National des Sciences Appliquées), Rennes, France |
| 1993 |
DEUG A (Diploma of Higher Education) University of the South Pacific, Tahiti, French Polynesia |
| 1991 |
Baccalauréat série E (A levels) Lycée technique du Taaone, Tahiti, French Polynesia |
Virtual reality and sports
I have to write this...
Synthetic humanoids animation
Real-time motion adaptation
These works developed during my PhD thesis deal with motion adaptation for hundreds of characters in real-time. It is based on a normalized representation of the skeleton (cf. Kulpa et al., Eurographics 2005). The kinematical and kinetical motion adaptation algorithm is an iterative process that verifies constraints while controlling the position of the center of mass. It adapts the movement at each time without taking future into account. It allows the use of these adaptations in real-time on hundreds of characters. This process is used in our studies in virtual reality.
|
Here is an example of the use of constraints of inequality. Indeed, the character on the left plays the original captured motion while the right one must preserve his arms above the table. |
|
|
|
 |
This example shows the adaptation of the same motion by using constraints placed on the skin of the forearms (red spheres). The targets of these constraints are placed on the table (yellow spheres). It illustrates that the placement of the constraints are not restricted to the skeleton. |
|
|
|
|
Two teams compete in a karate class. The characters in each team use the same movements. But to account for differences in morphology, the constraints are adjusted in real time. Thus, the movement of a character will automatically adapt to the movement of the other character that is itself altered by stress. This ability to interactively change the gesture of the two characters without prior knowledge of each movement is possible because the adaptive algorithm considers only the current posture. |
|
|
|
|
This example shows the interactive control of 40 characters in real time. The feet are fixed on the ground while the wrists are interactively controlled by the user. The first character shows the difference between the kinematical adaptation (in normal colors) and the combined kinematical and kinetical adaptations (in red). The combined adaptations are applied to the other characters. |
|
|
|
|
Only one motion is used to animate all the characters of this example but different constraints are used on their kicking foot. Even if the motion is modified, the center of mass is preserved at its original position in the motion. |
|
|
|
 |
Three different masses are added to the wrists of the characters. Kinematical and kinetical adaptations are applied to preserve the center of mass on the same vertical line. The feet are forced on the ground that moves in real-time. |
|
|
|
 |
This example shows the interactive control of the movement of a character. This control is achieved using only one reference movement (here that of a drunk person) and a destination that the character must reach (here represented as a vertical red line). The destination of the character is modified interactively by the character using the arrows of the keyboard. |
http://denzalabelleetoile.wix.com/chihuahua
|