Open Accessmammography classification model trained from image labels only
Author(s) -
Fredrik A. Dahl,
Matthew T. G. Holden,
Olav Brautaset,
Line Eikvil
Publication year - 2022
Publication title -
proceedings of the northern lights deep learning workshop
Language(s) - English
Resource type - Journals
ISSN - 2703-6928
DOI - 10.7557/18.6244
Subject(s) - computer science , artificial intelligence , annotation , pixel , mammography , pattern recognition (psychology) , population , machine learning , focus (optics) , norwegian , process (computing) , breast cancer , cancer , medicine , linguistics , philosophy , physics , environmental health , optics , operating system
The Cancer Registry of Norway organises a population-based breast cancer screening program, where 250 000 women participate each year. The interpretation of the screening mammograms is a manual process, but deep neural networks are showing potential in mammographic screening. Most methods focus on methods trained from pixel-level annotations, but these require expertise and are time-consuming to produce. Through the screenings, image level annotations are however readily available. In this work we present a few models trained from image level annotations from the Norwegian dataset: a holistic model, an attention model and an ensemble model. We compared their performance with that of pretrained models based on pixel-level annotations, trained on international datasets. From this we found that models trained on our local data with image-level annotation gave considerably better performance than the pretrained models from external data, although based on pixel-level annotations.