Tangible Interaction + Graphical Interpretation:
Bringing Clay Models to Life


Published at SIGGRAPH 2000:
Anderson, Frankel, Marks, Agarwala, Beardsley, Hodgins, Leigh, Ryall, Sullivan, and Yedidia, ``Tangible Interaction + Graphical Interpretation: A New Approach to 3D Modeling",  Proceedings of SIGGRAPH 2000, Annual Conference Series, July 2000.

Our vision for this project was to give children tangible access to 3D modeling using simple clay, and then to bring those clay models to life.  Children were asked to create clay models from a family of common types, such as a dog, person, car,  table, etc....  We then scanned these clay figures into 3D volumetric models.  Using a set of parameterized object templates which were optimized to best fit the volumetric model, I was able to identify which type of clay figure the child had made, and to parse the model into its constituent parts.  I then fed the descriptions of these parts into a motion controller which allowed us to generate realistic motion  for the child's clay model.  In this way, a clay dog made with the child's own hands could be brought to life in a 3D animation.  The top row of figures above show this process for a dog model, and the bottom row of figures shows the parses for a variety of clay models.

I developed all the algorithms for the volumetric model recognition, parsing, and animation.   This work formed half of the paper above, alongside another case study demonstrating simple tangible modeling coupled with interpretation of those models.  I did this work during a research internship with Joe Marks at MERL.

Aseem Agarwala