An Annotation Tool for Digital Restoration of Wall Paintings

Abstract

Antique paintings are essential to study and understand our past. Paintings, and specifically mural paintings, are delicate artworks that are affected by multiple deterioration conditions. Weathering and human interventions cause different damage problems, and physical and chemical changes degrade their visual color appearance. As a consequence, art historians and archaeologists require a huge effort to attempt to rebuild their original appearance. The annotation of digital images of the paintings is a valuable tool in this process. In this paper we analyze major requirements from art historians concerning the annotation of painting regions from the point of view of digital restoration. We also describe a tool prototype (based on TagLab) intended to facilitate the annotation and segmentation of mural paintings. The tool assists art historians in formulating multiple hypotheses on the original appearance by supporting multiple annotation layers for degradation and color, providing both hand-drawn and semi-automatic segmentation, and offering web-based dissemination and sharing of the annotations through the W3C Web Annotation Data Model.