The Shelburne Museum in Vermont is home to a renowned collection of American wildfowl and fish decoys. During renovation of the Dorset House where the decoys are usually on display, Nancy Ravenel, Objects Conservator, had the opportunity to examine some decoys more in depth. In the process, she explored the pros and cons of two type of three-dimensional x-radiography: computed tomography (CT) and volume rad tomosynthesis (VolumeRAD – a GE Healthcare trademark). Since the museum does not have its own radiography capabilities and is located in rural Vermont, there was no access to industrial imaging resources. Instead, Ravenel explored how best to maximize the capabilities from the local medical community through collaboration with the University of Vermont Medical Center.
For this exploration of radiographic techniques, the decoys proved to be excellent patients since they are somewhat simple in construction, yet personalized between makers and specific when used for hunting versus collecting. As an added bonus, they are easy to transport to the medical center. Ravenel used the Barnes swan as a case study while she looked for a maker’s mark at the head / neck joint.
Samuel Barnes,
Swan Decoy, c. 1890
Collection of Shelburne Museum, 1952-192.4
With the CT scan, Ravenel found that the metal elements cause flares, which can be distracting. Beam hardening on the image was also apparent. Since CT scanning requires specialized equipment, it is harder to schedule causing limited availability. On the other hand, CT data offers 360 degree data with options for viewing in a variety of ways. Examples of CT imaging on two ducks in the Shelburne collection can be viewed here https://youtu.be/FFjRmEat5xE and here https://youtu.be/bH3zEtzKRWs.
In contrast, the VolumeRAD technique captures data with the same equipment as standard radiography offering better accessibility. It also requires less radiation so there is less impact on the image from beam hardening. Cons to the technique include that the data is non-isotropic, the edges are not distinct, and there are fewer options for how the data is viewed. Ravenel also pointed out that it collects data of a short depth, so she has to identify where the imaging should take place, otherwise the results can be fuzzy. This can require some trial and error.
This image was taken at the University of Vermont Medical Center department of diagnostic radiology was part of a volume rad study of the joint between the neck and body of the decoy in order to locate a maker’s mark thought to be within the joint. The technique takes a series of images at set angles, thus avoiding the effect of the fasteners in the joint between the head and neck. The numeral “III” scratched into the joint is easier to see in this technique than it was in a standard posterior-anterior view radiograph.
Samuel Barnes,
Swan Decoy, c. 1890
Collection of Shelburne Museum, 1952-192.4
In the end, Ravenel felt that the VolumeRAD technique shows considerable promise and felt that she was better able to visualize the hollowing bit marks, dowels, and saw marks, which were all more distinct than in the CT scans. VolumeRAD, as a new technique, has considerable room for development and refinement.
An additional note beyond the presentation, there was some follow up discussion on viewing software. Ravenel noted in her presentation that she uses OsiriX, a DICOM viewer, for working with the data once back at the museum. An audience member pointed out that ImageJ is being widely used. Ravenel confirmed that she feels most comfortable with OsiriX and finds it to be more user friendly, while the audience member was quite happy with ImageJ and felt that it had deeper capabilities for the conservation community.
For more images of Shelburne decoys with radiographic images, visit their Flickr page here: https://www.flickr.com/photos/shelburnemuseum/albums/72157650406031226.
Skip to content
AIC Blog Archives: Conservators Converse
Former Blog of the American Institute for Conservation