In this work, we present a content-based image retrieval framework for the automatic recognition of archaeological artifacts from the collection of the Museum of Archaeology of the University of Catania (MAUC), using images acquired with a standard photographic camera. Compared to smartphone-based acquisition, camera imagery introduces additional challenges due to higher variability in optics, resolution, illumination, and acquisition parameters, making the retrieval task more complex. To evaluate the proposed framework under realistic and challenging conditions, a visual database of approximately 6,500 images was collected, capturing 35 artifacts from multiple viewpoints and under diverse lighting conditions. The large scale of the dataset further increases the difficulty of the retrieval task and provides a suitable case study to assess both robustness and scalability. We compare results obtained using two CNN-based feature extractors, assessing both recognition accuracy and computational requirements. Experimental results show that the proposed framework achieves high recognition performance while maintaining low computational cost, demonstrating that reliable artifact identification is feasible even with camera-based images and large visual databases.

Image Retrieval for Artifact Recognition Using a Photographing Camera: A Case Study

Allegra D.;Gueli A. M.;Stanco F.
2026-01-01

Abstract

In this work, we present a content-based image retrieval framework for the automatic recognition of archaeological artifacts from the collection of the Museum of Archaeology of the University of Catania (MAUC), using images acquired with a standard photographic camera. Compared to smartphone-based acquisition, camera imagery introduces additional challenges due to higher variability in optics, resolution, illumination, and acquisition parameters, making the retrieval task more complex. To evaluate the proposed framework under realistic and challenging conditions, a visual database of approximately 6,500 images was collected, capturing 35 artifacts from multiple viewpoints and under diverse lighting conditions. The large scale of the dataset further increases the difficulty of the retrieval task and provides a suitable case study to assess both robustness and scalability. We compare results obtained using two CNN-based feature extractors, assessing both recognition accuracy and computational requirements. Experimental results show that the proposed framework achieves high recognition performance while maintaining low computational cost, demonstrating that reliable artifact identification is feasible even with camera-based images and large visual databases.
2026
Computer Vision
Convolutional Neural Network
Cultural Heritage
Digital Humanities
Image Retrieval
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11769/738711
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact