Interpretable Classification of Levantine Ceramic Thin Sections via Neural Networks

2025-06-13

AI Summary

Addressing the time-intensive, expert-dependent nature of petrographic classification for Bronze-to-Iron Age ceramic thin sections from the Levant, this study proposes an interpretable deep learning framework. We employ transfer learning with ResNet-18 and Vision Transformer (ViT) architectures, and—novel in archaeological petrography—systematically integrate Grad-CAM and ViT’s self-attention mechanisms to visualize diagnostic textural and mineralogical features. The models achieve classification accuracies of 92.11% and 88.34%, respectively, with attention maps consistently highlighting quartz, feldspar, and other archaeologically significant mineral phases, closely aligning with expert petrographic interpretations. This approach bridges high predictive performance with decision transparency, advancing archaeometric practice toward reproducible, verifiable methodologies. It provides a robust, intelligent aid for ceramic provenance analysis and ancient technological reconstruction.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Machine Learning: Deep Neural Architectures and Foundation Models
• Data Mining & Knowledge Management: Representing, Reasoning, and Using Provenance, Trust

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Classification of ceramic thin sections is fundamental for understanding ancient pottery production techniques, provenance, and trade networks. Although effective, traditional petrographic analysis is time-consuming. This study explores the application of deep learning models, specifically Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs), as complementary tools to support the classification of Levantine ceramics based on their petrographic fabrics. A dataset of 1,424 thin section images from 178 ceramic samples belonging to several archaeological sites across the Levantine area, mostly from the Bronze Age, with few samples dating to the Iron Age, was used to train and evaluate these models. The results demonstrate that transfer learning significantly improves classification performance, with a ResNet18 model achieving 92.11% accuracy and a ViT reaching 88.34%. Explainability techniques, including Guided Grad-CAM and attention maps, were applied to interpret and visualize the models' decisions, revealing that both CNNs and ViTs successfully focus on key mineralogical features for the classification of the samples into their respective petrographic fabrics. These findings highlight the potential of explainable AI in archaeometric studies, providing a reproducible and efficient methodology for ceramic analysis while maintaining transparency in model decision-making.

Problem

• Classify Levantine ceramic thin sections using neural networks
• Improve classification accuracy with transfer learning techniques
• Interpret model decisions via explainability methods in archaeometry

Innovation

• Uses CNNs and Vision Transformers for classification
• Applies transfer learning to boost accuracy
• Employs explainability techniques for model decisions