Artificial Intelligence in Digital Forensic Pathology A Comprehensive Review of Deep Learning, Whole-Slide Imaging, and Explainable AI in Forensic Investigations

Abstract

The recent interaction between artificial intelligence (AI) and forensic pathology is transforming the study process of post-mortem evidence by medical examiners, coroners and pathologists. Digital forensic pathology, especially with the introduction of whole-slide imaging (WSI) and large-resolution autopsy imaging, has created huge transactions of rich visual data, which are becoming more susceptible to computational processing. This review systematically reviews AI usage focused on deep learning, convolutional neural network (CNNs), and transformer-based architecture on automated analysis of histopathological glass and virtual autopsy images and cause-of-death discrimination pattern recognition. We also discuss explainable AI (XAI) progress that makes the decisions of the algorithm explainable so that they can be used in the process of legal litigation, and ethical, law-related and regulatory issues of deploying AI-assisted diagnostics in medicolegal contexts. This review is based on more than 30 recent studies, which supports a systematic overview of the present state of affairs, as well as provides future perspectives of the development of the responsible use of AI in the field of forensic pathology.