Neural network for archaeological glyph detection

A novel approach leveraging convolutional neural networks (CNNs) significantly enhances the detection and classification of ancient glyphs in Sardinia's Neolithic Domus de Janas tombs, addressing challenges posed by environmental degradation. By preprocessing images to isolate glyphs from backgrounds, researchers mitigated distortions caused by lighting and wear, improving analysis accuracy. The study employed AlexNet, GoogLeNet, and EfficientNet models to categorize glyphs and identify geometric features, offering a more efficient method for preserving and studying these culturally significant artifacts.

Crisci, S., De Simone, V., Diana, A., & Zullo, F. (2025). Neural network for archaeological glyph detection. Intelligent Systems with Applications, 27, 200562. https://doi.org/10.1016/j.iswa.2025.200562