Laplacian Salience-Gated Feature Pyramid Network for Accurate Liver Vessel Segmentation.

Abstract

Liver vessels generated from computed tomography are usually pretty small, which poses major challenges for satisfactory vessel segmentation, including 1) the scarcity of high-quality and large-volume vessel masks, 2) the difficulty in capturing vessel-specific features, and 3) the heavily imbalanced distribution of vessels and liver tissues. To advance, a sophisticated model and an elaborated dataset have been built. The model has a newly conceived Laplacian salience filter that highlights vessel-like regions and suppresses other liver regions to shape the vessel-specific feature learning and to balance vessels against others. It is further coupled with a pyramid deep learning architecture to capture different levels of features, thus improving the feature formulation. Experiments show that this model markedly outperforms the state-of-the-art approaches, achieving a relative improvement of Dice score by at least 1.63% compared to the existing best model on available datasets. More promisingly, the averaged Dice score produced by the existing models on the newly constructed dataset is as high as 0.734±0.070 , which is at least 18.3% higher than that obtained from the existing best dataset under the same settings. These observations suggest that the proposed Laplacian salience, together with the elaborated dataset, can be helpful for liver vessel segmentation.