Simultaneous and automatic two surface detection of renal cortex in 3D CT images by enhanced sparse shape composition

摘要

Automatic organ localization plays a significant role in medical image segmentation. This paper introduces anovel approach for simultaneous and automatic two surface detection of renal cortex from contrast enhanced ab-dominal CT scans. The proposed framework is an integrated procedure consisting of three main parts: (ⅰ) cortexmodel training, both two shape variabilities are detected using principal components analysis from the manualannotation, and dual shape dictionaries and appearance dictionaries are constructed; (ⅱ) outer mesh reconstruc-tion, the initialized outer mesh is iteratively deformed to the target boundary; (ⅲ) inner mesh reconstruction,the inner mesh can be reconstructed using the same deformation coefficients and similarity transformation of theouter mesh with the inner mesh shape dictionary. Our method was validated on a clinical data set of 37 CT scansusing the leave-one-out cross validation strategy. The proposed method has improved the overall segmentationaccuracy of Dice similarity coefficient to 91.95%-3.15% for renal cortex segmentation.