Parameter constraints and ill-conditioning in tensor tomography reconstruction: a theoretical and numerical approach

Optics express +

|

Abstract

Tensor tomography is essential in material science and medical imaging for anisotropic, non-invasive internal structure analysis, especially for reconstructing anisotropic scattering properties beyond conventional tomography. It aids in studying material and tissue microstructures, revealing composition and mechanical properties. However, the mathematical representation of scattering tensors in reconstruction remains challenging, necessitating optimized acquisition parameters for accuracy. This work uses spherical harmonic and singular value decomposition to establish a theoretical upper limit of 15 parameters for accurately approximating the scattering function in anisotropic X-ray dark-field tomography (AXDT). Our findings highlight the importance of selecting optimal parameters for reliable reconstruction and provide a condition number for acquisition scheme stability analysis. These insights improve tensor tomography system design, enhancing efficiency in medical and material diagnostics.