Unraphael

The study of Raphael's paintings reveals numerous faithful reproductions, highlighting both the prolific nature of his workshop's output and the enduring appeal of his designs. Traditionally, comparing these reproductions has been a labour-intensive process, constrained by manual examination and limited case studies.

The software called Unraphael introduces a digital workflow tool designed to enhance the analysis of paintings and their reproductions by uncovering the various transfer methods employed in creating copies. Focusing on Raphael’s oeuvre, this software covers a comprehensive pipeline utilizing advanced computer vision techniques to precisely extract and compare figure outlines across reproductions. This pipeline includes preprocessing, background removal and individual figure extraction, image alignment, and clustering based on structural similarity indices and brushstroke signatures. This enables the differentiation of various transfer techniques with high accuracy.

This software demonstrates that computer vision can substantially refine the examination of differences and similarities between original works and their reproductions. It not only offers new insights into Raphael's artistic practices but also enhances our understanding of copying techniques from the 15th and 16th centuries.