The primary goal of this project is to develop a non-invasive, machine learning-enabled software solution that can accurately detect and quantify both surface and internal features in pharmaceutical tablets using three-dimensional X-ray microscopy (XRM) imaging. By integrating 3D image processing techniques with artificial intelligence (AI), this project aims to enhance quality control processes in drug product manufacturing. While machine learning has already been applied in this field, existing methods typically analyze 3D images slice by slice, treating them as a stack of 2D images rather than as a single three-dimensional structure. The novelty of this project lies in the development of an approach that can analyze tablet microstructures in true three-dimensional space, simultaneously considering spatial relationships across all dimensions. There has been little to no research on machine learning models capable of performing 3D instance segmentation for pharmaceutical tablets in this manner. To achieve this, the project will focus on developing a machine learning-based 3D instance segmentation model to directly process volumetric XRM data. These models will be optimized for high accuracy and precision in detecting tablet features, enabling more comprehensive defect detection and material characterization. The developed AI models will then be integrated into DigiM’s existing software, ensuring seamless adoption into current pharmaceutical workflows. The software package will be validated using tablets supplied by DigiM, with performance benchmarked against traditional methods. By capturing and analyzing pharmaceutical tablets in three dimensions, the project aims to provide insights into tablet microstructures, ultimately optimizing manufacturing processes.