利用可適性彈性蛇變曲線模型從事核磁共振影像之腦部區域分割

Abstract

醫學影像分割是個一直以來被廣泛研究的領域。在這篇論文中，我們提出一個新的全自動參數化主動輪廓模型 (Parametric Active Contour) 彈性蛇變曲線來進行腦部區域的擷取。我們所提出的方法其架構主要分成兩個部分: 影像前處理和影像分割。首先，利用模糊可能性分群模型 (FPCM) 對腦部影像進行分類，其分類後的標記影像為後續輪廓初始化步驟的主要參考依據。在第二個部分裡，藉由參考前一階段所計算出來的標記影像，將輪廓初始化在腦部區域的外圍，並在彈性力的驅使下逐步變形。彈性蛇變曲線模型利用可適性法向力使得輪廓向內收斂以擷取腦部區域的邊界。我們將此方法應用在 T1-權重的核磁共振影像上的腦部分割，並在和其他著名的腦部區域分割方法比較之下，顯示我們所提出的方法具有良好的分割結果且具有廣泛的醫學影像分割應用潛能。

Brain image segmentation, also known as skull-stripping, has been the focus of a wide variety of research in recent years. This paper proposes a new automatic parametric active contour model (snake) for brain image extraction. The proposed framework consists of two stages: image preprocessing and image segmentation. First, the fuzzy possibilistic c-means (FPCM) is used for voxel clustering, which provides a labeled image for the following contour initialization. At the second stage, the contour is initialized outside the brain surface based on the result of the FPCM and evolves under the guidance of the balloon force. The balloon snake model drives the contour with an adaptive inward normal force to capture the boundary of the brain. The proposed algorithm is evaluated by segmenting a number of T1-weighted magnetic resonance images. Experimental results and comparisons with other existing approaches show the effectiveness of this new scheme and potential applications in a wide variety of brain image segmentation.