3D Shape Analysis for Archaeology

Archaeology is rapidly approaching an impasse in its ability to handle the overwhelming amount and complexity of the data generated by archaeological research. In this paper, we describe some results of our efforts in developing automatic shape analysis techniques for supporting several fundamental tasks in archaeology. These tasks include documentation, looking for corollaries, and restoration. We assume that the input to our algorithms is 3D scans of archaeological artifacts. Given these scans, we describe three techniques of documentation, for producing 3D visual descriptions of the scans, which are all non-photorealistic. We then proceed to explain our algorithm for partial similarity of 3D shapes, which can be used to query databases of shape, searching for corollaries. Finally, within restoration, we describe our results for digital completion of broken 3D shapes, for reconstruction of 3D shapes based on their line drawing illustrations, and for restoration of colors on 3D objects. We believe that when digital archaeological reports will spread around the globe and scanned 3D representations replace the 2D ones, our methods will not only accelerate, but also improve the results obtained by the current manual procedures.