Enhancing Ceramic Petrography through Deep Learning
Author(s): Ashley Rutkoski; Nicolas Gauthier; Neill Wallis; Andrea Torvinen; Ann Cordell
Year: 2024
Summary
This is an abstract from the "Ceramic Petrographers in the Americas: Recent Research and Methodological Advances" session, at the 89th annual meeting of the Society for American Archaeology.
Clay recipes reveal information about the local geology and the inclusion of different additives that make up a vessel, which in turn reflects the social, environmental, and technological context of ceramic manufacture. Ceramic petrography has long been instrumental in shedding light on key manufacturing techniques, identifying unique mineralogical signatures, and assessing patterns of cultural exchange among diverse communities. However, traditional methods used to characterize and quantify petrographic thin sections can be labor-intensive and time-consuming, making it difficult to meet the increasing demand for large comparative datasets. Here, we present recently developed machine-learning algorithms to enhance the efficiency, accuracy, and accessibility of these intricate analyses. We review the basic steps for conducting these approaches and illustrate their practical application for being incorporated into a petrographer’s standard toolkit. Streamlining petrographic workflows through machine-learning techniques will open new avenues for the quantitative and qualitative assessment of ceramic technology and its broader implications for past societies.
Cite this Record
Enhancing Ceramic Petrography through Deep Learning. Ashley Rutkoski, Nicolas Gauthier, Neill Wallis, Andrea Torvinen, Ann Cordell. Presented at The 89th Annual Meeting of the Society for American Archaeology. 2024 ( tDAR id: 498406)
This Resource is Part of the Following Collections
Keywords
General
Communities of Practice
•
digital archaeology
Geographic Keywords
Multi-regional/comparative
Record Identifiers
Abstract Id(s): 40419.0