Enhancing level set brain tumor segmentation using fuzzy shape prior information and deep learning

摘要

Magnetic resonance imaging (MRI) brain tumor segmentation is a crucial taskfor clinical treatment. However, it is challenging owing to variations in type,size, and location of tumors. In addition, anatomical variation in individuals,intensity non-uniformity, and noises adversely affect brain tumor segmentation.To address these challenges, an automatic region-based brain tumor segmentationapproach is presented in this paper which combines fuzzy shapeprior term and deep learning. We define a new energy function in which anAdaptively Regularized Kernel-Based Fuzzy C-Means (ARKFCM) Clusteringalgorithm is utilized for inferring the shape of the tumor to be embedded intothe level set method. In this way, some shortcomings of traditional level setmethods such as contour leakage and shrinkage have been eliminated. Moreover,a fully automated method is achieved by using U-Net to obtain the initialcontour, reducing sensitivity to initial contour selection. The proposed methodis validated on the BraTS 2017 benchmark dataset for brain tumor segmentation.Average values of Dice, Jaccard, Sensitivity and specificity are 0.93± 0.03, 0.86 ± 0.06, 0.95 ± 0.04, and 0.99 ± 0.003, respectively. Experimentalresults indicate that the proposed method outperforms the other state-of-theartmethods in brain tumor segmentation.