Research Abstracts - 2006

Acquiring Deformable Objects

Daniel Vlasic & Jovan Popović

Abstract

We are interested in extracting the 3D shape of deformable objects such as hands, faces, or full bodies. In addition, we would like to recover a deformation model for such an object.

One of our envisioned approaches entails observing the object (e.g. with a video camera) while it deforms, and aggregating all the surface information over time. For example, if someone waves a hand in front of the camera, clenches the fist, bends the fingers, our system would learn the shape of the hand as well as the fact that fingers can bend at the joints.

Another direction we have explored concentrates more on the full-body motions. We have devised a system that uses ultrasonic time-of-flight measurements to triangulate 3D positions of markers placed on the body. The figures show reconstructions of markers placed on the arms. With many such markers we can acquire much of the body's deformation.

As an end goal, we would like to be able to acquire the full human body as it moves and deforms. This would enable us to enrich existing motion databases and help us gain more insight in human motion.

References

[1] Peter Sand, Leonard McMillan, Jovan Popović. Continuous Capture of Skin Deformation. ACM Transactions on Graphics 22(3), 2003, pages 578-586.

[2] Robert W. Sumner, Jovan Popović. Deformation transfer for triangle meshes. ACM Transactions on Graphics 23(3) 2004, pages 399-405.

[3] Robert W. Sumner, Matthias Zwicker, Craig Gotsman, Jovan Popović. Mesh-based Inverse Kinematics. ACM Transactions on Graphics 24(3), 2005, pages 488-495.