Published in:
01-08-2019 | Endometrial Cancer | News
In this issue
Published in:
Virchows Archiv
|
Issue 2/2019
Login to get access
Excerpt
This issue opens with a review on a hot topic: the use of artificial intelligence in pathology. Komura and Ishikawa (
https://doi.org/10.1007/s00428-019-02594-w) discuss machine learning techniques, notably those using deep neural (convolutional) and generative adversarial networks. The authors argue that machine learning applied to histopathological images requires special approaches, because it is somewhat of a black box and most pathologists would hesitate to make a (machine generated) diagnosis without understanding on which elements the machine diagnosis is based. Many pathologists tend to be skeptical of the capacity of this technology to classify histopathological images, exactly for this reason. The authors briefly review some applications and approaches to solve existing problems. …