Signed Laplacian Deep Learning with Adversarial Augmentation for Improved Mammography Diagnosis
This addresses the problem of improving mammography diagnosis for breast cancer detection, which is incremental as it builds on existing deep learning methods with novel regularization and augmentation.
The paper tackled the problem of computer-aided breast cancer diagnosis in mammography, which is limited by inadequate data and similarity between benign and cancerous masses, by proposing a signed graph regularized deep neural network with adversarial augmentation, and the result showed that the DiagNet framework outperformed the state-of-the-art in breast mass diagnosis.
Computer-aided breast cancer diagnosis in mammography is limited by inadequate data and the similarity between benign and cancerous masses. To address this, we propose a signed graph regularized deep neural network with adversarial augmentation, named \textsc{DiagNet}. Firstly, we use adversarial learning to generate positive and negative mass-contained mammograms for each mass class. After that, a signed similarity graph is built upon the expanded data to further highlight the discrimination. Finally, a deep convolutional neural network is trained by jointly optimizing the signed graph regularization and classification loss. Experiments show that the \textsc{DiagNet} framework outperforms the state-of-the-art in breast mass diagnosis in mammography.