Graphics and Archaeology: Interpreting the Past
Snowbird, Utah
20-23 May 2000
Presentation:
- The Digital Michelangelo, Forma Urbis, and Cuneiform Tablet Projects
Presenters:
- Marc Levoy
Presentation Resources:
Summary:
Marc's presentation looked at some of the details of various recent projects concerning:
- 3D scanning of large statues
- digitizing light fields
- solving 3D jigsaw puzzles
- visualizing inscribed surfaces
The 3D scanning of large statues involved a custom camera and trusswork. It involved calibrated and uncalibrated movments of a triangular laser scanning system, resolving data through strip scans that were integrated in a subsequent step. The impressive scans of Michelangelo's David resulted in 2 billion primitives and 32 Gigabytes of data alone.
While the engineering details are presented in a much better fashion in Marc's papers and weblinks, some interesting highlights of the technology include:
- a color processing pipeline that allowed for the compensation of ambient, shadowed, or specular pixels (thus capturing color data that could more accurately be used in synthetically lit situations)
- applications of alternative rendering algorithms to the surface topology of the model (such as the probe sphere / accessibility rendering technique, much like rubbing a dark material into crevices and then wiping it away) to reveal insights not easily achieved in conventional light
- high-resolution scanning of the work, St. Matthew, revealed insights within Michelangelo's scuplting technique by giving access to fine chip and tooth detail of the chisel that are not readily visible on the actual sculpture
- ability to view the object from impossible angles (given its orientation and display in the gallery) - thus realizing that the statue was cleverly designed for two primary viewing angles (profile and quarter)
- it was also possible to ascertain that Michelangelo anticipated the viewer's angle of view and probable light conditions - example discovered was examination of David's brow line and its appearance under a variety of lighting conditions
- when building the gantry, it was discovered that the statue is 17' high (not 14' as reported in all history books)
- besides engineering issues, it was also important to ensure that collisions with the scanned objects were avoided and survivable
Marc continued on to light field rendering. It is a form of image-based rendering, capable of creating new views by rebinning [interpolation] of original views. Light fields begin as an array of images. The advantages included the lack of need of a 3D model, less computation than rendering a model, and rendering cost is independent of scene complexity. The disadvantages are fixed lighting, static scene geometry, and the need to stay outside of the convex hull of the object. The plan for the project was to image the Medici Chapel. Complications with the site (falling sections of roof) prevented the completion of this effort.
Advantages for the use of digitized models such as these include:
- unique views of these statues
- virtual museums
- physical replicas
- permanent archives
- geometric calculations
- projection of images onto statues
The Forma Urbis Romae Project is a plaster model of anicent Rome. Some 1,163 fragments exist of this great map - only about 15% of the actual map. The paper and website cover this project in excellent detail and were not covered in depth in this presentation.
In the closing parts of his presentation, Marc touched upon the visualization of cuneiform tablets. A photograph can only show one side of a tablet at a time (since the writing continues around the object). Using various visualization and projection techniques (displacement mapping, accessibility shading, and a projection unwrapping algorithm), Marc showed some examples of how tablets could be accurately viewed (all surfaces) all at once.
Conclusions / Issues:
- Datasets like these are huge - simplification or adaptive viewing tools are key in accessing this information effectively
- Intellectual property rights were a major issue - although this was an experiment of the technology, there were additional socio-political ramifications around control of the likenesses - not to be underestimated.