Global guidance network for breast lesion segmentation in ultrasound images

•First, we present a CNN (denoted as GG-Net) with a global guidance block (GGB) to aggregate non-local features in both spatial and channel domains under the guidance of multi-layer integrated features for learning a powerful non-local contextual information.•Second, we develop a breast lesion boundary detection (BD) module in shallow CNN layers to embed additional boundary maps of breast lesions for obtaining the segmentation result with high-quality boundaries.•Third, the experimental results on two ultrasound breast lesion datasets show that our network outperforms the state-of-the-art medical image segmentation methods on breast lesion segmentation.•Moreover, we also show the application of our network on the ultrasound prostate segmentation, where our network obtains satisfactory performance.