Evaluation of Annotation Ambiguity in Common Supervised Machine Learning Classification Approaches for Cultural Heritage

To date, the semantic classification of heritage assets is a widespread practice in digital heritage documentation processes. The technique involves the use of Artificial Intelligence (AI) to analyze unstructured 3D survey data, such as point clouds or meshes, via the semi-automatic identification of various architectural elements, including building components (Croce et al. in Sensors 23:2497, 2023 [1]), decay patterns (Barba et al., Representation Challenges New Frontiers of AR and AI Research for Cultural Heritage and Innovative Design. FrancoAngeli, pp. 171–178, 2021 [2]) and material maps (Schönfelder et al. in Autom. Constr. 152, 2023 [3]), based on a small amount of sample data.