Researchers have developed a new method for analyzing craniosynostosis severity using statistical shape models. The method, called TRACE, can work with imperfect surface scans from 3D photographs taken in clinical settings. These scans often contain artifacts like hair, clothing, and scanner noise that can corrupt the analysis. TRACE uses a template-constrained approach to predict corresponding head-surface control points and deform a clean template mesh into a subject-specific head reconstruction. The method has been tested with different backbones and shown to improve surface sampling, topology preservation, and shape-model quality compared to prior methods. --- Why it matters: This matters because it provides a scalable foundation for photograph-based craniosynostosis shape analysis, which could reduce the need for radiation-heavy computed tomography scans. It also demonstrates a framework that can be applied to other artifact-contaminated surface scans with an appropriate clean template. Source: https://arxiv.org/abs/2608.22131