Graphics and Archaeology: Interpreting the Past
Snowbird, Utah
20-23 May 2000
Presentation:
- Adaptively Sampled Distance Fields
Presenters:
- Alyn Rockwood
Presentation Resources:
Summary:
Alyn presented results of ongoing work at the MERL lab in Boston. He started his presentation with an introduction - "the untold part of the story". He has a passion for history and it is the interpretation of negative space (that which is not written) that can bring great enjoyment. Consider, for example, a challenge made to Hemingway to write a short story in only six words. Hemingway answered - "For Sale: Baby Shoes - never used." It is the human element that fills in the story.
Adaptively Sampled Distance Fields (ADF) are a graphics data structure that can unify many forms of computer graphics representations. Every point in space holds a distance to the object in question. It is like the water level in a terrain (where the actual water level is the object itself). Numerically, you are always going down hill towards the object (waterlevel) with the distances either being outside, on, or inside the object itself.
The datastructure itself is represented as an octree. It is raytraced to create a rendering and the data structure is directly evaluated as the ray traverses it. Modeling operations such as booleans are easily represented even for the most complex object. In fact, Alyn demonstrated that the complexity of objects (such as a fractal sponge) easily compressed resource usage over that required to present the surfaces with polygons.
Alyn showed examples of unique objects with a variety of complexities all rendered from this datastructure. They included the organic forms of molecules, milled surfaces, a high order parametric surface with a complex ornament channeled across its surface. This representation offers easy capture of both organic and parametric surfaces.
Future work will be around rendering, modeling, deformations, data acquisition and conversion, and many more tasks. It is all highly relevant to the techniques being discussed at this archaeology symposium.
This work will be presented at the SIGGRAPH 2000 conference in New Orleans.
Conclusions / Issues:
- There exists a "chicken and egg" dilemma around this technology - databases are not available yet, but do we wait for the tools to be finished first?
- There are potentials for this work to be applied to the paint and light analysis work being developed by the collaboration of the University of Bristol, DERA, and the National Gallery.