0
$\begingroup$

What is the difference between a histopathological image and a natural image when training a neural network?

$\endgroup$
1
$\begingroup$

Histopathological images compared to natural images difference so, that in histopathological images the image needed to be modeled may contain millions of pixels, whereas a recognition from natural image that this is a dog / house / certain person needs significantly less information to be extracted from the picture.

Histopathological images use WSIs (Whole slide image) which contain the part of tissue as a whole. It needs to be split to 256 x 256 pixel patches and those are used for ROI hunting (Region of Interest) and other analysis one by one.

Histopathological images (WSI) are quite rare nowadays and hold privately. More open data would rise the accuracy of machine learning and increase the amount of results that could be found from the data.

Complete source:

https://arxiv.org/pdf/1709.00786v1.pdf

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.