Clifford代数几何不变量3D医学图像配准的方法

Approach for 3D Medical Image Registration Based on Clifford Algebra Geometrical Invariance

就3D医学图像配准数据量大、计算复杂度高、配准精度低的问题,提出一种基于Clifford代数几何不变量的配准方法,以实现人头颅部3D医学图像配准。提出配准所需的Clifford代数几何不变量及其Clifford代数方程算式,并构造适合于该几何参考轴旋转的Clifford几何旋转算子,利用所求的最大、最小值对应的Clifford几何不变量建立Clifford旋转算子,对浮动影像数据实现几何变换,以达到配准的结果。配准实验中对两个世界著名的3D医学数据集进行了测试,结果表明:该方法计算简单,几何意义直观,配准精度高,执行效率高,并且通过轴线变换不易陷入配准过程的局部极值点。

Considering that 3D medical image registration has the problems of huge data,high computational complexity and low registration precision,this paper proposed a registration method based on Clifford algebra geometric invariants to realize 3D medical image registration of skull part for human,proposed Clifford algebra geometric invariants and Clifford algebra equation formulas needed by registration,constructed Clifford geometric rotation operator which is fit for rotation for the geometric reference axis,established the rotation composite operator using the corresponding Clifford geometric invariants obtained by the maximum and minimum values.To realize registration,geometric transformation was made for the floating image data.Two famous 3D medical data were tested in registration experiments.The experimental result indicates that this method has the advantages of simple calculation method,intuitive geometric meaning,high registration precision,high execution efficiency,and also not easily falls into the local extreme value in the registration process.