Computer Aided Analysis of Undrawings in IR-Reflectograms

Interdisciplinary projects between the field of art history and digital image analysis brought new aspects in both of the fields. While art historians benefit from new objective analysis methods and improved efficiency, for technicians a new field of application and research was opened. This proposal presents a computer aided system to analyze underdrawings in medieval paintings. Underdrawings are the basic concept of an artist when he starts the creation of his work of art. A systematic analysis of underdrawings, starting with medieval paintings, will bring insights to the practice in painting schools, drawing media, drawing tools, and drawing styles. This new experience is important for both, art historians and conservators. The technical or engineering goal is to set up the environment for a systematic analysis of underdrawings. To visualize an underdrawing, which is usually hidden by covering pigment layers and therefore invisible to the observer in the visible light spectrum, specific infrared sensors will be used. Infrared sensors which are sensitive especially in the spectral range from 1.000 nm to 2.200 nm will be produce digital images of the underdrawing (infrared reflectograms). The goal from the research point of view is to investigate image processing methods for analysis of underdrawings. In contrast to other infrared projects our goal is not only to digitize, visualize, and improve images of underdrawings but to analyze the structure of the underdrawings with image processing and pattern recognition methods. The system will allow to segment underdrawings from the background, analyze strokes in the drawings with respect to drawing tool and drawing material used, and it will further allow to differentiate between freehand drawings and drawings created by the help of stencils. The acquisition system must be able to provide images with a resolution high enough to be able to differentiate between different drawing tools. The systematic investigation will be supported by the installation of a database, that will consist of art historical information about the objects and results of the automated analysis process. Querying the database will allow to get an overview over the large amount of data. An additional visualization tool supports the analysis process.

Recent developments in computer vision provide powerful tools for the examination and classification of data of our cultural heritage. New camera hardware allows new insights into cultural heritage, especially if infrared cameras are concerned, since they allow the study of structures that are visually hidden. The methods developed in the field of digital image analysis for this project assist art experts in analyzing painted work of art in order to have a more accurate, efficient and objective analysis of the underdrawings. The data acquisition is based on conventional images in the visible light, images in the near infrared range and on multi-spectral images in the near infrared range for a multi-spectral analysis of the data. The methods developed allow an elimination of the painting cracks, an automated segmentation of strokes in the underdrawing as well as the analysis of underdrawing strokes and the color pigments in the color layers of the painting. The segmentation of underdrawing strokes extracts strokes from the background of the image. The goal is to extract structural information from the arrangement of the strokes in underdrawings in order to obtain artist specific characteristics that allow for a stylistic analysis of drawing. Further more the segmentation is a pre-processing step for the analysis of the strokes with regard to drawing tool and material. The recognition of the drawing tool and material is a considerably aspect of art experts in analyzing underdrawings. Therefore we perform texture and boundary analysis algorithms in order to extract textural and boundary features of the different strokes. The features gained are the data for pattern recognition algorithms in order to identify the drawing tool and material. A focal point in this interdisciplinary project was to extend the collaboration between art historians and computer scientists. Therefore we studied in this project the collection from the Museum of Medieval Art, Austrian Gallery Belvedere. Approximately 1000 IR-images from objects in the Austrian Gallery have been taken with a CCD- camera. The collection of images gives a first overview about the underdrawings from this period. In cooperation with the Austrian Gallery a new formulation of the museum catalog is in preparation. Our contribution will be on the one hand the IR images taken in this project. On the other hand we will utilize our developed tools to provide additional aspects of the underdrawing.