基于三维图像的大脑病灶显示和分析

Brain lesion display and analysis based on 3D rendering

针对头部各组织在灰度分布上存在复杂的粘连和交错叠加现象,提出在管状加强滤波算法中增加体素亮度调和参数并与管状控制的其他参数一起共同控制管状响应值的产生,该管状加强滤波算法和其他相关切割算法通过采用GPU编程方式实现,并对脑部肿瘤的提取和显示技术进行了探讨。实验结果表明,本文方法对于脑部组织和相关病灶的分割提供可视的实时显示,对脑部疾病的微创手术和介入治疗提供图像导视。

In view of the complex adhesion and overlapping of brain tissues in gray distribution,a novel vessel enhancement filtering algorithm combining original vessel control parameters and newly added voxel luminance harmonization parameter is proposed to produce vessel response values.The vessel enhancement filtering algorithm and relative segmentation methods are implemented by GPU.The technologies for brain tumour extraction and display are also explored.The experiments show that the proposed method can provide virtual reality for the segmentations of brain tissues and relative lesions,and offer image guidance for minimally invasive surgery and interventional therapy for brain diseases.