A Framework for Integrating Domain Knowledge & Deep Learning for 3D Shape Analysis of Lithic Fragments
Year: 2024
Type: conference-paper
Abstract: The increasing volume of data available for archaeological analysis requires automation,and advances in machine learning and artificial intelligence are being deployed to meetthese needs.Deep learning approaches, in particular have been very successful in manyareas. These methods, however, are something of a ‘black box’ – while they successfullymap from input data to the desired output, they are difficult to interpret. Related to thisis the fact that they perform best when trained ‘end-to-end’ which makes incorporatingexpert knowledge into them difficult. In this paper we propose a new direction, incorporating more explicit reasoning intodeep learning systems for fine-grained 3D shape analysis. Our target application isthe analysis of lithic flakes from pre-contact Māori stone tool manufacture. As well ashaving significant archaeological interest, this application exemplifies some of the challengesof fine-grained shape analysis. While our work is at an early stage, we believethat the overall approach may be of interest to other members of the community asthey explore how to use deep learning to solve archaeological problems.
Authors: Mills, Steven, Das, Nirmal, O'Regan, Gerard, Arun, Lana, Chakraborti, Tapabrata, Walter, Richard
Institutions: University of Otago, University College London
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6949410499
FWCI:
Cites:
Cited by:
Related to: