SlicerMorph photogrammetry: an open-source photogrammetry workflow for reconstructing 3D models

High-fidelity three-dimensional (3D) models of skeletal specimens underpin many ecological and evolutionary analyses. Here we present a fully open pipeline inside the 3D Slicer platform that couples automatic image masking by the Segment Anything Model (SAM) with surface reconstruction by the OpenDroneMap (NodeODM) engine, all wrapped in a user-friendly extension. To test accuracy, we photographed 14 mountain-beaver skulls, reconstructed 3D models with the new pipeline and with our previous workflow and compared each model to its micro-CT reference using mean surface distance, root mean square error (RMSE), Hausdorff, and Chamfer metrics. Our improved pipeline that integrates masking to the model reconstructed lowered mean distance and RMSE by 10-15% across specimens and reduced visible artefacts around thin elements such as zygomatic arches; Hausdorff distance changed little, indicating that gains were global rather than confined to outliers. Our new extension provides a convenient workflow that integrates masking, scaling, and reconstruction under one interface.