Keywords: Bilateral mammogram, asymmetry attention, breast cancer
Abstract: Radiologists commonly compare the bilateral mammograms to detect asymmetric abnormalities. While fibroglandular tissue is normally quite symmetrically distributed, lesions in one breast and will only rarely have a counterpart in the corresponding area of the opposite breast. Motivated by this experience, we explore a model that can learn to detect asymmetrical information from bilateral mammograms and then find the abnormal areas, similar to what a radiologist does. This can increase model performance and interpretability. We evaluate the proposed methods on the popular INBreast dataset and show improved performance in abnormal classification and weakly supervised segmentation tasks.
Registration: I acknowledge that acceptance of this work at MIDL requires at least one of the authors to register and present the work during the conference.
Authorship: I confirm that I am the author of this work and that it has not been submitted to another publication before.
Paper Type: novel methodological ideas without extensive validation
Primary Subject Area: Detection and Diagnosis
Secondary Subject Area: Application: Radiology
Confidentiality And Author Instructions: I read the call for papers and author instructions. I acknowledge that exceeding the page limit and/or altering the latex template can result in desk rejection.