Characterisation of organs in a post-mortem computed tomography database using machine learning

In this thesis, the effects of a wide range of decedent-presentations on an automated labelling pipeline for analysing post-mortem computed tomography (CT) scans are investigated. This document is organised into three investigations: development of a machine learning (ML) pipeline applied to forensic data; the development of a tool for generation of synthetic trabecular structures; and, the quantification of the effect of artefacts on image quality caused by dense structures in CT scans.