The Batch Artifact Scanning Protocol: A new method using computed tomography (CT) to rapidly create three-dimensional models of objects from large collections en masse

TL;DR

The paper presents the Batch Artifact Scanning Protocol, a CT-based workflow to rapidly generate high-quality 3D models from large artifact collections, addressing the bottlenecks of traditional single-object scanning. It details a three-stage process—preparation, scanning, and post-processing—with automated segmentation and surfacing implemented in Python (dicom_firstpass.py, dicom_refine.py, surface.py) and using a 2000 HU threshold for bone fragments, producing meshes via Marching Cubes. In a case study with 2,474 ungulate bone fragments across 329 packets, the method achieved about 10.75 hours of scanning and roughly 3 hours of post-processing, translating to around 4.38 seconds per fragment for surfacing and substantial time savings over photogrammetry. The approach enables scalable data generation, supports data sharing and machine-learning workflows, and has broad implications for archaeology, taphonomy, education, and public-facing museums by enabling rapid, non-destructive 3D modeling of large collections.

Abstract

Within anthropology, the use of three-dimensional (3D) imaging has become increasingly common and widespread since it broadens the available avenues for addressing a wide range of key anthropological issues. The ease with which 3D models can be generated and shared has major impact on research, cultural heritage, education, science communication, and public engagement, as well as contributing to the preservation of the physical specimens and archiving collections in widely accessible data bases. Current scanning protocols have the ability to create the required research quality 3D models; however, they tend to be time and labor intensive and not practical when working with large collections. Here we describe a streamlined Batch Artifact Scanning Protocol to rapidly create 3D models using a medical CT scanner. While this method can be used on a variety of material types, we have, for specificity, applied our protocol to a large collection of experimentally broken ungulate limb bones. By employing the Batch Artifact Scanning Protocol, we were able to efficiently create 3D models of 2,474 bone fragments at a rate of less than 4 minutes per specimen.

Figures (11)

• Figure 1: Supplies for scanning Shown here are the supplies we used for scanning which include polyethylene foam, a cutting mat, painter's tape, a hot glue gun with glue sticks, a utility knife, a smart phone for taking photographs for the purpose of documentation, and a laptop to create the .csv companion file. While this setup was chosen with the protection of the scanned objects in mind, it should be noted that any packaging material can be used as long as its density is discernibly different from the target object during scanning.
• Figure 2: Fragment placement The fragments should not overlap in the $x-$ or $y-$ directions. This ensures that the automated segmentation can properly separate the fragments within the scan data into individual models for surfacing. The $x-axis$ is the view from the side of the scanning bed. The $y-axis$ is the bird's eye view of the scanning bed.
• Figure 3: Documenting specimens for scanning Here we provide an example of how to complete the formatted $.csv$ so that the segmentation and surfacing algorithms will function properly. The first column indicates the date (YYYYMMDD), the second column indicates the packet number for that date, the third column indicates the direction the code should read the .csv file, the fourth column indicates whether or not the scan was mirrored, and the remaining columns indicate the specimen labels. The third and fourth columns are there to mitigate the need to resurface the scan should it have been oriented improperly on the scanning bed.
• Figure 4: Fragment layout Here we offer images of the various stages of the packaging process. They are lettered according to the order of operation within the protocol. We took a photograph of the layout of the fragments for the entire package (A). Photographs were taken of individual fragments such that we could clearly see the labels (B). If the fragment was not directly labeled we included the labeled bag in the photograph (C). We traced the fragments using a sharpie (D) and then used the outline to cut out sections in the foam to encase the fragments (E). Packets were wrapped in tape for additional protection during transport (F).
• Figure 5: Example of package label Pictured here is an example of how we labeled the package indicating how to orient the package on the scanning bed and as a cross-reference for the .csv file for that scan package.