Geometric Analysis for Classification and Reassembly of Broken Bones
This project applies modern geometric and data analytical tools to the study and automated reassembly of broken bone fragments, in order to determine the agent of breakage (whether carnivore, hominin, or other taphonomic agent), improve taxonomic identifications, and better understand site formation processes through spatial analysis of refits. Both controlled samples of ungulate bones that have been broken by known agents and field samples found in Dmanisi, Republic of Georgia, will be analyzed. The anthropological implications are expected to impact the current understanding of early human prehistory, culture, and origins. The mathematical techniques to be developed are based on invariant signatures and moving frames, including differential invariants, integral invariants, and invariant histograms, and will incorporate (semi-)supervised learning algorithms on graphs, spectral clustering, symmetry-based numerical approximation schemes, and the reconstruction of three-dimensional objects from camera projections. This project is being led by Katrina Yezzi-Woodley, Jeff Calder, Peter Olver, and Martha Tappen.
Funding for this project comes from the NSF Grant DMS-1816917, the University of Minnesota Graduate Research Partnership Program (GRPP), Anthropology Department block grants, the University of Minnesota Thesis Research Travel Grant, University of Minnesota Undergraduate Research Opportunities Program (UROP), the University of Minnesota Grant in Aid (2017-2019), and the NSF BIGDATA IIS-1837992.
Digitization of Manual Methods in Lithic and Zooarchaeological Analysis
Through this project we are developing a digital goniometer that, with limited user supervision, can rapidly extract angle data from 3D models of fragmentary bone and lithic flakes. Powerful computational algorithms applied to 3D models can bring traditional manual methods for data collection to the digital world. Data extraction becomes faster, easily replicable, and more mathematically robust. The end goal is to develop a digital goniometer that easy to use and freely available for use by other researchers. This project is being led by Katrina Yezzi-Woodley, Jeff Calder, Peter Olver, Annie Melton, and Gilbert Tostevin.
Funding for this project comes from the NSF Grant DMS-1816917 and Anthropology Department block grants.
Utilizing Refitted Core Sequences in the Quantitative Assessment of Cultural Transmission
This project seeks to understand cultural transmission processes among prehistoric populations, focusing specifically on refitted lithic core nodules. We are exploring new techniques for quantifying cultural transmission among Paleolithic populations. Specifically looking at refitted core reduction nodules, this project plans on combining known cultural transmission proxy variable data with the specific sequence data provided by the refits to achieve a higher resolution understanding of variability in stone tool production among prehistoric populations. To do so, known time-series mathematical applications are being explored as a possible approach. The overall goals will be to address concerning the presence or absence of cultural continuity in the Levantine Middle to Upper Paleolithic and the degree of production variability within and among lithic assemblages. Refitted core sequences from Boker Tachtit, a transitional Middle to Upper Paleolithic site in Israel, and Taramsa-1, a Middle Paleolithic site in Egypt, will serve as the sample for this project. This project is led by Annie Melton, Gil Tostevin, Jeff Calder, and Brendan Cook.