中文摘要：

在计算机断层造影增强图像（CTA）体数据中,骨骼是血管三维可视化的重要障碍。针对CTA体数据的去骨问题,提出一种基于特征和层间相关性的CTA体数据自动去骨算法。首先,利用CTA体数据的CT值,对骨骼和血管进行粗分割;然后,利用骨骼和血管的空间形态特征,对过分割的血管区域进行还原;最后,在三维空间模式下,对残余骨骼体素和与骨骼粘连的血管区域进行特征识别和层间相关性处理,在不需要用户手动参与的情况下,实现骨骼的全自动去除。对49套CTA体数据进行实验,结果表明：该算法具有很强的鲁棒性,实现速度快,对患者的伤害小,完全满足临床诊断中的血管三维可视化要求。

英文摘要：

Bone creates a major barrier in visualizing the 3 D blood vessel tree.The method is based on feature and interlayer correlation from computed tomography angiography（CTA） volume data and is proposed in the paper.Firstly,bone and blood vessel was segmented roughly by the value of CTA volume data.Then,the over-segmentation blood vessel region was restored by the feature of spatial shape.In the end,based on structure feature and interlayer correlation,the remaining bone and vessels connected to the bone was processed automatically without manual intervention.The experimental results on 49 CTA volume data indicate that the proposed method is robust for bone removal and the computation time is less.Moreover,it is considered to be safe for patients and can satisfy with the 3 D visualization demands in clinical diagnosis.