In this talk, Lieve Watteeuw showed images produced by the Reflectance Imaging for Cultural Heritage (RICH) project and demonstrated the functionality of the Mini-Dome module. I was excited to see this presentation after reading advertisements for the “New Bownde” Conference at the Folger Shakespeare Library last year, in which the RICH project and Mini-Dome were featured. Fortunately, for those unable to attend the presentation, extensive documentation about the project is available online through the project’s webpage and blog.
The Mini-Dome module is a small, hemispherical-shaped imaging device that is tethered to a laptop. The module creates dynamic digital images through polynomial texture mapping, a technique commonly referred to as Reflectance Transformation Imaging (RTI). This technique involves taking a series of images from a fixed camera position, while changing the angle of lighting, in order to reveal the surface of an object. The original dome was created in 2005 at Katholieke Universiteit Leuven for reading cuneiform tablets, but has since been used to image bindings, illumination, wax seals, cuir bouilli, and other cultural objects. The current module is equipped with a single 28 MP digital camera and 260 white LED lights to capture a total of 260 images in approximately four minutes. Watteeuw showed a video of the Mini-dome in action during her presentation, but readers can view a similar video here.
After capture, Watteeuw described how the images are processed by software and interpreted by seven dynamic filters. She demonstrated some of these filters, including sharpen, shading, generate models, line drawings, and sketch. Using the mouse or interface buttons, the user can zoom in, drag the image, or change the direction of lighting in real-time. For those that would like to experience the software interface, a web viewer is available here. Watteeuw reports that the software suite can also export to 3D shaded or rendered modes.
Watteeuw’s presentation included a demonstration of a measurement tool built into the image processing software. The tool can be used to measure the distance between two points or generate a height map for a portion of the object. This blog post includes an image of a height map created on the blind-tooled surface of a leather binding. Watteeuw explained that by scaling the image, the measurements can be accurate to 10 microns.
Watteeuw’s presentation included several examples of how the Mini-Dome could be used to learn more about the production of a binding. Images of a late 15th century book of hours were manipulated with filters to show tool marks on the uncovered wooden boards, providing evidence of how the boards were shaped and the lacing of the sewing supports. A second example showed a 16th century book (pictured above and described here), in which the binder scored a vertical line in the leather to align a large, central impression. Watteeuw described instances in which previously unknown marks or designs were revealed by manipulating the filters or direction of the lighting, such as three leaves emerging from an emblem design, or shallow impressions from a decorative roll being more clearly defined. This tool could be quite useful for identifying individual finishing tools and documenting how they changed or became damaged over years of use.
In addition to leather bindings, Watteeuw shared images of Belgian damask silk, remains of a ribbon, and embroidered bindings from the Folger Shakespeare Library. Once again, the dynamic filters in the software suite were applied in order to enhance details of the objects. The “sketch” tool provides clear images of weaving and embroidery patterns and could be very useful to textile historians and conservators. The measurement tool could also be used to gather data on the thickness of threads or cord used to construct the object.
The RICH project will continue until 2015 and additional investigations are already underway. Watteeuw reports that the topographic data from an object is exportable into spreadsheet form. Engineers on the team are currently exporting high points of objects scanned to create a large data set for further analysis. Additional projects include the development of optical character recognition (OCR) for specific tool shapes or patterns on bindings. Testing is proceeding on a smaller “micro-dome” (pictured above) that is constructed in two pieces so that it can be placed inside the opening of a book to capture images of the gutter or surface of a page. Watteeuw described a student research project currently in progress to measure sewing in manuscript textblocks.
Two questions were asked by audience members following the presentation. The first individual asked if a database of the existing images is available. Watteeuw answered that an open access database of all images captured would be ideal; however, since this is a research project using prototypes, the team is collaborating with institutions to link with existing databases. A second audience member asked if any attempts had been made to identify tools of various workshops. Watteeuw replied that a corpus was needed before any comparisons could be made.
Advanced imaging technologies, such as RTI, offer tremendous opportunities for the study of cultural objects and for digital libraries in general. The RICH project has produced a suite of tools that could be used by scholars and practicing conservators to gain a better understanding of an object’s composition and production. Wider use of devices such as the Mini-Dome in imaging collections of note and greater access to the software suite is required in order to exploit the full potential of the technology.
A fantastic article Henry, I’ve just shared it on my Google+ page! I had never heard of polynomial texture mapping before, I guess it’s just a matter of time before it’s somehow built into smartphones?! A fantastic advancement for the book arts, especially so I think for restoration specialists. I found a few other resources on hp’s website for anyone who’s looking: http://www.hpl.hp.com/research/ptm/
Keep up the good work!
Paul