SlicerMorph Photogrammetry: An Open-Source Photogrammetry Workflow for Reconstructing 3D Models

1. ContextAccurate three-dimensional (3D) models of skeletal and other biological specimens are crucial for ecological and evolutionary research and teaching. Here, we present a streamlined, open-source workflow for 3D photogrammetry that allows researchers to construct 3D models using the 3D Slicer and the SlicerMorph ecosystem. 2. Objectives and MethodsWe present an updated photogrammetry pipeline within the 3D Slicer ecosystem that combines advanced image masking via Segment Anything Model (SAM) with the open-source ODM reconstruction engine. Our approach systematically reduces artifacts in delicate cranial regions by automating background removal. We validate this pipeline against micro-CT references using multiple rodent skull specimens, quantifying accuracy through mean distance, root mean square error (RMSE), Hausdorff distance, and Chamfer distance. 3. ResultsOur method consistently reduces average surface error by 10-15% compared to a previous open-source photogrammetry workflow. This gain is particularly pronounced around thin or fenestrated anatomical structures such as zygomatic arches, where the previous workflow created the most noise. Although occasional increases in Hausdorff distance reflect localized photography gaps, improved mean distance and RMSE underscore improved overall fidelity. 4. Main ConclusionsThe 3D Slicer Photogrammetry extension provides a robust, fully open-source solution for generating high-quality 3D reconstructions from photographs. The streamlined pipeline lowers technical barriers to collecting reliable 3D data and fosters reproducible research across diverse biological collections.