Locally-run interpret-able breast cancer diagnosis from histology images
Date
2021-03-31
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The study furthers artificial intelligence/machine Deep Learning in medical diagnostics, and works to explain decisions behind model predictions and to make the algorithm transparent. It adds another level of visual explanation for the algorithm’s decision making process. However, to improve the algorithm’s predicting power, careful considerations need to be paid to how histology images are pre-processed. To solve issues of privacy, the study uses solutions that allow such models to run locally. Convolutional Neural Networks (CNNs) is a class of deep learning algorithm trained on large volumes of labelled radiological images.
Description
item.page.type
Conference Paper
item.page.format
Keywords
ARTIFICIAL INTELLIGENCE, AI4D, DIAGNOSIS, MEDICAL IMAGING, MACHINE LEARNING, RADIOLOGY, BREAST CANCER, HISTOLOGY, ALGORITHM, SOUTH OF SAHARA