Archaeology with AI

Sabrano analytical and documentary departments are backed by artificial intelligence and deep learning models, enabling easy integration of local knowledge.

Five strategic objectives of the World Heritage Convention are our guidelines: Credibility, Conservation, Capacity Building, Communication, and Communities.

Working with textual and imagery archives, we rely on fast processing of material through image recognition and large language model frameworks that are constantly perfected.

Integration with LiDAR mapping involves:

-Data layers in QGIS or ArcGIS, with

• Overlay detection of features on construction plans
• Export of 3D models and maps for site evaluation

-Field Validation and ground-truth verification at flagged locations

Sabrano AI framework has been trained on a large number of imagery and textual datasets in the medical field, biochemistry, and physics. Application in archaeology and cultural heritage studies is a natural continuation of this remarkable technology.

Parallels between Archaeology and Life Sciences include:

-visible surface; land-cover or cell structure which is scanned with a microscope or airborne camera (plane)

-documents; different written formats/languages, hand-written pages, recognition and text interpretations

_

Deep Learning Models:

U-Net (semantic segmentation for detecting subtle features in remote sensing data)

Mask R-CNN (Instance segmentation for identifying distinct archaeological features)

ResNet/CNN (Classification models for patch-based analysis