Convolutional neural networks on postmodern histology organ recognition

Abstract

Convolutional neural network (CNN) has advanced in recent years and translated from research into medical practice. Research on CNNs in forensic/post mortem pathology is almost exclusive to post mortem computed tomography, despite the wealth of research into CNNs in surgical/anatomical histopathology. This study was carried out to investigate whether CNNs are able to recognise different organs on histology slides. This study compared four CNNs commonly used in surgical/anatomical histopathology to identify microscopic images of brain, heart, kidney, lung and liver. One of the CNNs used (InceptionResNet v2) was able to show a >95% accuracy in classifying the organs. The result of this study is promising and demonstrates that CNN technology has potential applications as a screening and probably a computer assisted diagnostics tool in forensic/post mortem histopathology.