Using Computer Vision and Deep Learning Algorithms to Predict Pottery Types: An Example Using Ancestral Pueblo Pottery from the Central Mesa Verde Region

Author(s): Dylan Schwindt; Kari Schleher; Michelle Turner; Grant Coffey; Benjamin Bellorado

Year: 2021

Summary

This is an abstract from the "SAA 2021: General Sessions" session, at the 86th annual meeting of the Society for American Archaeology.

Computer vision, machine learning, and artificial intelligence techniques have made much progress in the past several years. Cloud computing has rendered these tools more accessible than ever to researchers in a wide range of fields. Here we explore applications of these models to classify Ancestral Pueblo pottery types in the central Mesa Verde region of southwestern Colorado. We explore a range of models, from deep learning models based solely on image analysis to models that include imagery combined with simple, user-provided observations to predict pottery typology. We also compare deep learning models that predict typology directly with more traditional models that use attributes to predict topology and models that combine both approaches. Finally, we discuss the most promising predictive models for public education products about archaeology and ancient Indigenous technologies.

Cite this Record

Using Computer Vision and Deep Learning Algorithms to Predict Pottery Types: An Example Using Ancestral Pueblo Pottery from the Central Mesa Verde Region. Dylan Schwindt, Kari Schleher, Michelle Turner, Grant Coffey, Benjamin Bellorado. Presented at The 86th Annual Meeting of the Society for American Archaeology. 2021 ( tDAR id: 467732)

This Resource is Part of the Following Collections

Keywords

General
Ancestral Pueblo • digital archaeology • Machine Learning • Material Culture and Technology

Geographic Keywords
North America: Southwest United States

Spatial Coverage

min long: -124.365; min lat: 25.958 ; max long: -93.428; max lat: 41.902 ;

Record Identifiers

Abstract Id(s): 33361