The article explores a method for early detection of breast cancer using mammography in a realistic hospital setting. It challenges the constraints of existing methods, which require annotations for individual images or regions of interest (ROIs), and a fixed number of images per patient. Instead, it proposes a weakly supervised learning setting where labels are available per case, but not for individual images or ROIs. The researchers propose a two-level multi-instance learning (MIL) approach for breast cancer prediction. The study shows that two-level MIL can be applied in realistic clinical settings where only case labels, and a variable number of images per patient are available.
Publication date: 20 Oct 2023
Project Page: Not provided
Paper: https://arxiv.org/pdf/2310.12677
