TomoSuitePY: A Pythonic Workflow for Noisy and Sparse Angle Tomography

Advancements in machine learning have introduced innovative approaches for analyzing and enhancing tomographic datasets. However, many of these neural networks pose challenges for non-technical users, limiting their practical application at high-speed synchrotron tomography instruments. This manuscript introduces TomoSuitePY, a Python-based module that streamlines multiple neural networks into a user-friendly workflow for tomographic denoising and upsampling sparse angles. Within this framework, we propose a novel approach to address sparse angle datasets by interpolating frames between projections. To validate this method, we present various performance metrics applied to an ex-situ study of a neural network on a cathode material, specifically LiNi0.8Mn0.1Co0.1O2.