Cascaded context enhancement network for automatic skin lesion segmentation

Skin lesion segmentation is an important step for automatic melanoma diagnosis. Due to the non-negligible diversity of lesions from different patients, extracting powerful context fine-grained semantic still challenging today. Although deep convolutional neural networks (CNNs) have made significant improvements on skin segmentation, they often fail reserve spatial details and long-range dependencies due consecutive convolution striding pooling operations inside CNNs. In this paper, we formulate a cascaded enhancement network segmentation. A new aggregation (CCA) module with gate-based information integration approach proposed sequentially selectively aggregate original image multi-level features encoder sub-network. The generated further utilized guide discriminative extraction by designed context-guided local affinity (CGL) module. Furthermore, auxiliary loss added CCA refining prediction. our work, evaluate four public dermoscopy datasets. method achieves Jaccard Index (JA) 87.1%, 80.3%, 84.3%, 86.6% ISIC-2016, ISIC-2017, ISIC-2018, PH2 datasets, which show highly competitive performance other state-of-the-art models respectively. • Propose generates richer context. extracts fine feature global Achieving performances