Deep Multi-instance Networks with Sparse Label Assignment for Whole Mammogram Classification

TL;DR

This paper introduces end-to-end deep multi-instance networks for whole mammogram classification that eliminate the need for costly manual annotations, demonstrating robustness on the INbreast dataset.

Contribution

It proposes three novel deep multi-instance learning schemes for mammogram classification, reducing annotation costs and improving robustness over previous methods.

Findings

Deep networks outperform segmentation-based methods

Three multi-instance schemes are effective for whole mammogram classification

Robust performance demonstrated on INbreast dataset

Abstract

Mammogram classification is directly related to computer-aided diagnosis of breast cancer. Traditional methods requires great effort to annotate the training data by costly manual labeling and specialized computational models to detect these annotations during test. Inspired by the success of using deep convolutional features for natural image analysis and multi-instance learning for labeling a set of instances/patches, we propose end-to-end trained deep multi-instance networks for mass classification based on whole mammogram without the aforementioned costly need to annotate the training data. We explore three different schemes to construct deep multi-instance networks for whole mammogram classification. Experimental results on the INbreast dataset demonstrate the robustness of proposed deep networks compared to previous work using segmentation and detection annotations in the training.