See-Through Museum project: a complete imaging pipeline for cultural and natural heritage

The See-Through Museum Project was launched in May of 2016 with an idea of bringing together researchers with an expertise in advanced x-ray tomography, structured light scanning and data analysis with conservators and restorers of several prominent Dutch and Chinese museums. The ultimate aim of the project is to create an integrated data acquisition, visualization and analysis pipeline allowing collection of the most relevant 3D data for a wide range of heritage artifacts. We believe that the state-of-the-art imaging technologies will in the near future enable museums to have comprehensive digitized collections. These virtual collections will be accessible by audiences remarkably wider than an audience of a typical modern museum. Besides that, the abundance of imaging data will allow solving numerous research questions related to artifact design and conservation. However, currently, such a breakthrough is constrained by technical challenges in the 3D scanning process, which we address in our work. Our focus lies within the field of tomographic methods (x-ray, neutron or visible light), as they allow to acquire a true 3D representation of an artifact. Some challenges of applying tomography to museum artifacts are related to their great variability. Variability of material composition makes it hard to choose optimal acquisition settings and often leads to errors in reconstructed images. Variability of sizes and form factors sometimes makes the scanning process unfeasible. On the other hand, combination of the data produced by different imaging modalities and analysis of that data require expert knowledge of a technical scientist and can often not be done by a cultural heritage expert. We would like to present a number of showcases where we were able to find solutions to some of the challenges mentioned above. These showcases include imaging and data analysis of samples of Burgess Shale fossils, Chinese carved ivory balls, pottery, Dutch shipwreck textiles and dendrochronology of wooden specimens. In each case, we have optimized the imaging conditions, algorithms used for tomographic reconstruction and developed tools that allow for quantitative data analysis. We hope, that ultimately this work will result in an integrated imaging pipeline that will cover the whole process of digitization of heritage artifacts, starting from optimization of raw data acquisition to analysis and visualization.