Digital Shape Matching Reunites History's Lost Fragments

Many objects entered museum collections at times when excavation and collecting practices were very different from today. In the past, excavated objects were often divided between institutions around the world, and display was prioritized over documentation. Over time, connections between pieces were lost. As a result, museums around the world hold remarkable artifacts whose backstories are thin, fragmentary or missing altogether.

In Egyptology, a quiet crisis exists. Museums around the world, from Europe to America, display beautiful ancient artifacts in glass cases. Many of these objects were collected in the 19th and early 20th centuries. However, their histories are often missing. You can see a painted mummy's foot case or a golden mask, but we do not know where they came from. They are beautiful, but historically they are silent.

During earlier eras of archaeology, excavation and collecting practices were very different. Objects from a single dig were often divided among different museums. Careful display was often more important than careful record-keeping. Over the decades, the connections between related pieces were lost. As a result, museum artifacts now have stories that are thin, broken, or gone entirely.

Today, archaeologists regularly find broken fragments that were once part of larger objects. Sometimes, these new fragments might share the same underlying shape with objects already in museums. For example, a mummy's foot case in a museum and a newly found piece might have been made in the same mold, sharing a consistent three-dimensional form. Yet time, distance, and a lack of records keep them apart.

Traditionally, determining if a fragment matches a museum object relied on a researcher's visual judgment and old, often incomplete, records. It was not based on a quantitative, or number-based, comparison of shape. This gap between field discoveries and museum collections has been a lasting challenge. My research asks a simple question: Can we use digital tools to test if fragments and museum objects might be related? If we can, we could recover parts of their lost histories.

Archaeology is, by nature, a study of fragments. Objects break, decay, or are moved over centuries. For a long time, archaeologists have used visual inspection, stylistic comparison, and written records to suggest links between fragments and objects. These methods are essential, but they have limits. Visual judgments can be subjective, and old records are often incomplete or inconsistent.

Because of this, many possible links between excavated material and museum artifacts have remained only guesses. An object in a museum may look complete, but its history is fragmented. Without a way to test these relationships systematically, many fragments are seen as secondary or unimportant.

More than a century ago, the famous archaeologist Flinders Petrie argued that an object's true value lies not in its beauty, but in the information it carries. He believed an ordinary fragment with a known history could be more important than a finely made object without one. Today, digital tools are giving archaeologists new ways to put that idea into practice.

One of these key tools is 3D scanning. Using portable scanners, it is now possible to capture the full surface shape of an object with high precision. This can be done without touching or damaging the artifact. Every curve, contour, and variation in thickness can be recorded digitally.

Once scanned, an artifact becomes more than a picture. It becomes data: a detailed digital model. This model can be rotated, measured, compared, and analyzed. Importantly, this process is noninvasive. Fragile objects do not need to be moved, taken apart, or physically tested.

For archaeologists and museum curators, this opens new doors. Objects held in different institutions, or fragments stored in excavation archives, can be compared digitally. This can happen even if the original objects never leave their storage locations.

Scanning is only the first step. The real challenge is in comparison. Instead of just asking if two pieces look similar, computational shape analysis lets researchers ask a more precise question: How similar are their shapes?

In simple terms, the computer compares the geometry of two surfaces. It looks at curvature, thickness, and spatial relationships, measuring how closely one surface matches another. It is like comparing a geometric fingerprint.

This approach does not replace expert judgment. Instead, it supports it by providing measurable evidence. This evidence can confirm, refine, or challenge a researcher's visual impression, allowing archaeologists to move from intuition to testing.

In a recent study, I applied these methods to Graeco-Roman Egyptian funerary artifacts. These objects were made of cartonnage, a material made from linen, plaster, and paint.

I created high-resolution 3D scans of excavated cartonnage fragments and compared them with an intact funerary mask held in a museum. The goal was not to physically rebuild the object, but to test if their shapes were compatible in a meaningful way.

The comparison focused on three-dimensional geometry rather than painted decoration. This is key because cartonnage masks were often shaped in molds. If two objects were formed in the same mold, they can share highly consistent curvature and thickness patterns, even if their painted surfaces look different.

I used a distance-mapping approach called deviation mapping. After aligning the 3D model of an excavated fragment to the corresponding region of the intact museum object, the algorithm calculates the distances between the two surfaces at thousands of points. Areas where the distances were consistently small are geometrically very similar. Areas with consistently larger distances indicate that the fragment’s shape diverges from the reference surface.

I used a method called deviation mapping. After aligning the 3D model of an excavated fragment to the matching area of the intact museum mask, the computer calculates the distances between the two surfaces at thousands of points. Areas where the distances are consistently very small are geometrically very similar, while areas with larger distances show where the fragment's shape differs from the reference surface.

In this case, the surfaces matched closely. Differences were generally less than one millimeter. This level of agreement is consistent with production in the same mold and is not just a coincidental visual resemblance.

The most important result was not just finding a single "match," but the ability to evaluate relationships in a transparent and repeatable way using shared digital evidence.

One powerful aspect of this approach is that it works across great distances. Researchers can easily share digital models and compare fragments and objects held in different institutions without moving fragile artifacts. Excavation archives, museum collections, and research labs can begin to speak the same digital language, reconnecting evidence that has been separated by geography and history for a very long time.

The work described here is part of my recent CRAFT Project. It does not use artificial intelligence (AI) or machine learning; instead, it relies on computer-based shape comparison and careful interpretation of measurements. However, it is part of a larger movement in heritage research.

Globally, researchers and institutions are starting to combine 3D scanning with machine learning, exploring collections in new ways. For example, the European Union-funded RePAIR project uses AI and robotics to help put together shattered archaeological artifacts, and major institutions like the Smithsonian are testing AI-driven analysis of large 3D collections.

Together, these projects point to a future where digital tools play a more active role in how museums and archaeologists understand the past.

Digital archaeology is sometimes linked to flashy 3D reconstructions or virtual displays, but its deeper value is found elsewhere. By giving fragments a new analytical role, digital methods allow archaeologists to recover relationships once thought lost forever.

New digital methods are bringing fresh energy to an old archaeological principle. Modest fragments can carry great significance when they help clarify an object's origins and its lost context. In this way, they can finally help an artifact find its way back home.