Learning from adversarial medical images for X-ray breast mass segmentation.

• We propose a novel framework for breast mass segmentation associated with small labeled datasets based on ACP-based parallel medical imaging framework. • We introduce a cGAN-based model to generate X-ray breast mass images with pixel-wise masks in the step of designing artificial medical imaging system. • We introduce an improved U-net model for mass segmentation in the step of computational experiments. • Thorough computational experiments on a public mammogram database of INbreast and a private database show that our proposed method can generate realistic lesion images with precise masks and further enhance the performance of deep segmentation network. Simulation of diverse lesions in images is proposed and applied to overcome the scarcity of labeled data, which has hindered the application of deep learning in medical imaging. However, most of current studies focus on generating samples with class labels for classification and detection rather than segmentation, because generating images with precise masks remains a challenge. Therefore, we aim to generate realistic medical images with precise masks for improving lesion segmentation in mammagrams. In this paper, we propose a new framework for improving X-ray breast mass segmentation performance aided by generated adversarial lesion images with precise masks. Firstly, we introduce a conditional generative adversarial network (cGAN) to learn the distribution of real mass images as well as a mapping between images and corresponding segmentation masks. Subsequently, a number of lesion images are generated from various binary input masks using the generator in the trained cGAN. Then the generated adversarial samples are concatenated with original samples to produce a dataset with increased diversity. Furthermore, we introduce an improved U-net and train it on the previous augmented dataset for breast mass segmentation. To demonstrate the effectiveness of our proposed method, we conduct experiments on publicly available mammogram database of INbreast and a private database provided by Nanfang Hospital in China. Experimental results show that an improvement up to 7% in Jaccard index can be achieved over the same model trained on original real lesion images. Our proposed method can be viewed as one of the first steps toward generating realistic X-ray breast mass images with masks for precise segmentation. [ABSTRACT FROM AUTHOR]