A Marker-Based Registration Approach for the Fusion of Ultrasound and 3D Surface Scans

Abstract

The fusion of ultrasound (US) imaging with threedimensional (3D) surface scanning enables the combination of internal anatomical and mechanical tissue properties with external body geometry and is particularly relevant for residual-limb assessment in prosthetic care. Accurate alignment of these modalities remains challenging due to differences in imaging principles, coordinate systems, and data representations. This work presents a marker-based registration workflow for integrating B-mode US, shear wave elastography, and 3D structured-light surface scans. Markers were localized in both modalities and the resulting corresponding coordinate pairs were used to compute a rigid point-based transformation aligning the US coordinate system with the 3D surface model. Validation on a phantom, a healthy subject, and a transfemoral amputee yielded mean fiducial registration error values ranging from 4.47 to 5.26 mm and enabled fused 3D visualization of internal tissue information within the surface model. These results demonstrate the technical feasibility of spatially consistent multimodal data fusion for residual-limb assessment.