Enhancing manuscript analysis with reflectance transformation imaging: a user study on interactive visualization options

Reflectance Transformation Imaging (RTI) is a well-established technique to improve the visualization of cultural heritage objects. Recent advancements have been achieved from the technical point of view, with novel interpolation techniques improving the quality of reconstructed highlights and shadows and algorithms for feature enhancement. The effectiveness of the relighting and enhancement methods proposed is, however, difficult to assess in practice, as it requires experts’ feedback on domain-specific data. This work aims at providing hints on the optimization of the effectiveness of the use of RTI for studying palimpsests, i.e., old documents in which writing has been removed, covered, or replaced by new writing. Leveraging a custom web visualization tool developed with an open-source library, we performed a user evaluation with field experts. We compared three relighting models differing in the ability to reproduce highfrequency effects and shadows and a series of image processing tools designed to enhance the visualization experience. We collected experts’ feedback with a properly designed questionnaire, including ratings of the usefulness of the techniques for the specific task and generic questions on user preferences. The survey highlighted that an application allowing image processing and relighting within the same interface is considered very important for the study of manuscripts. It is important to consider user feedback, as their interests and field of study guide their attention to specific features. Therefore, applications that allow users to choose between different relighting algorithms might be the most effective approach.