An improved region growing method for scalp and skull extraction based on mathematical morphology operations

Abstract

In this paper, we present a novel region growing method combined with mathematical morphology for scalp and skull extraction in T1-weighted brain MR images. First, a grayscale dilation operation is used in 2D brain MR image. Second, the scalp and skull portions are extracted by using a region growing method with intensity threshold. Third, the scalp and skull mask is optimized by using a closing operation. Finally, the mask model is reconstructed in three-dimensional space. The results of the proposed method and the region growing with intensity threshold on MRI data acquired from five persons were compared with manual segmentation. The mean similarity index for the scalp and skull extracted regions was equal to 96.2% in the proposed method. The experimental results indicate that the proposed method can extracted the scalp and skull accurately and effectively.