Active Shape Dictionary for Automatic Segmentation of Pathological Lung in Low-dose CT Image

摘要

Accurate lung segmentation is of great significance in clinical application. However, it is still a challengingtask due to its complex structures, pathological changes, individual differences and low image quality. In thepaper, a novel shape dictionary-based approach, named active shape dictionary, is introduced to automaticallydelineate pathological lungs from clinical 3D CT images. The active shape dictionary improves sparse shapecomposition in eigenvector space to effectively reduce local shape reconstruction error. The proposedframework makes the shape model to be iteratively deformed to target boundary with discriminativeappearance dictionary learning and gradient vector flow to drive the landmarks. The proposed algorithm istested on 40 3D low-dose CT images with lung tumors. Compared to state-of-the-art methods, the proposedapproach can robustly and accurately detect pathological lung surface.