What secrets can tomography uncover from an ancient vase?

July 07, 2025

Ancient pottery is like a time machine to the past – pottery relics give us insights into the cultures of the past. Most often, shards or fragments are found, rather than complete objects. Sometimes, nearly complete objects are found but are fragile and contain many cracks. But what exactly can X-ray CT reveal about an ancient vase? Let’s take a look.

The vase under investigation here was scanned by Diondo GmbH. Diondo is a leading CT system manufacturer and is one of Dragonfly’s sales partners. The vase is shown in Figure 1 in a photograph inside the CT system, and in a 3D CT view using Dragonfly 3D World. Figure 2 shows a cross sectional CT view indicating main features visible, a related slice video is available for further viewing. The vase is in private ownership (thank you to Adrian Tröscher for sharing) and its origins have been identified with inputs from Mr Mannack. It originates from southern Italy ca. 900-700 BC and is a special type of ancient Greek vase referred to as a Stamnos.

Figure 1 Ancient vase front view photograph (left), and CT 3D view (right).
Figure 2: Cross sectional view showing the internal details with some annotations indicating cracks, image artifacts and bright and dense areas related to density variations in the sample.

Full slice video of a sample.

Looking a bit closer (in this case with Dragonfly’s cross-hairs and oblique views), it is possible to identify some glue inside the cracks. In Figure 3 this is shown in quad-view with glue visible as a slightly darker shade of grey than the surrounding vase material. This glue is what keeps the vase together today.

Figure 3: Dragonfly quad-view with cross-hairs indicating the glue holding the majority of the vase together

In this vase it would be useful to visualize the cracks in 3D. However, the grey values in the image vary considerably, with varying density regions of the vase in addition to image artifacts, making it impossible to use “regular” thresholding operations to segment different parts of the image. Dragonfly’s deep learning comes in very useful here to segment all cracks and voids despite variations of brightness of other materials. The end result is shown in Figure 4, colored in yellow.

Figure 4: Yellow colored visualization of all cracks – segmented by deep learning.

An interesting capability of CT imaging is the possibility to re-build a broken object. Let’s imagine one of the pieces of the vase was fully broken off, if both are scanned, the parts can be put back together virtually. This is shown in Figure 5 where the central triangular fragment is removed and then replaced by 3D registration tools in Dragonfly, like a 3D jig-saw puzzle.

Figure 5: Demonstration of a virtual 3D jig-saw puzzle by replacing a broken fragment and putting it back in its correct 3D location. Image on the left shows the vase with fragment separate. On the right is shown the re-built vase.

So what does all of this mean? These are some take home messages:

  • CT is well suited to the non-destructive inspection of internal details of valuable and fragile objects like ancient pottery
  • It is possible to image the glue between the fragments and see this in cross-sectional images
  • It is possible to segment the cracks using deep learning
  • And it is possible to put together parts like a virtual 3D puzzle in case parts are broken off
  • And finally: if you have a broken vase, you can scan the parts and put them together virtually with Dragonfly 3D World

Full 3D CT animation of an ancient pottery vase

Did you enjoy this blog story? Get in touch to feature your unique application here in future blog posts.

Try Dragonfly 3D World

Get a Free Trial

More Real World Applications

Read Application Notes
Back to overview