Enhancing Archaeological Classification with Machine Learning: The Lincoln Pottery Works Collection

Summary

This is an abstract from the "SAA 2025: Individual Abstracts" session, at the 90th annual meeting of the Society for American Archaeology.

A significant challenge that many archaeological projects face is how to consistently and efficiently identify and classify artifacts, while processing a large corpus of data amassed through fieldwork. This is particularly poignant in smaller institutions that may lack the resources to utilize new and powerful machine learning techniques. In this poster, we propose a novel methodology that leverages machine learning techniques for the automatic identification and classification of artifacts.

Our case study is based on the Lincoln Pottery Works (LPW) Archaeological Collection. The LPW was a stoneware factory in Lincoln, Nebraska, operating from 1880 to 1903, which produced utilitarian, domestic wares. The LPW collection's substantial size of over 14,000 objects has made it impractical to manually produce a comprehensive and detailed catalogue of its contents.

We have developed a process that trains a machine learning model to identify and classify objects based on their specific attributes from both 2D images and 3D models of the artifacts within the collection. By creating a scalable and accessible framework, we aim to democratize the use of advanced machine learning approaches, enabling resource-constrained institutions to enhance their artifact analysis and preservation efforts.

This Resource is Part of the Following Collections

• Society for American Archaeology 90th Annual Meeting, Denver, CO (2025)   •
• SAA 2025: Individual Abstracts

Keywords

Record Identifiers

Abstract Id(s): 53066