改进Demons算法的非刚性医学图像配准

Non-rigid medical image registration based on improved Demons algorithm

非刚性配准是医学图像处理的一个重要的研究方向。基于光流场模型的Demons算法由于仅依赖图像灰度梯度使图像变形,当缺乏梯度信息时图像的变形方向不能确定,因而容易造成误配准,且该算法只适合于单模态图像配准。本文针对最大互信息配准方法在多模态刚性配准中的成功应用,提出了一种可用于多模态图像配准的改进Demons算法。该方法在原有驱动图像变形力的基础上,增加两幅图像间互信息对当前变换的梯度作为附加力作用,使浮动图像向两图像间互信息增大的方向变形,正确地配准图像。为避免陷入局部极值并提高算法的运行速度,该方法在多分辨率策略下实现。使用单模态、多模态图像分别进行实验来验证此算法,并与原始Demons算法进行比较,实验表明,该方法能够快速地产生准确的配准变换。

Non-rigid registration is one of the important research issues in medical image processing field. An intensity-based automatic deformable image registration algorithm, known as the “Demons” algorithm, only depends on the image intensity gradients of the reference image to drive the floating image transform, it easily results in wrong registration when gradients are small, even being zeros. Moreover, this algorithm is not fit for the registration of multi-modality images. So, an improved “Demons” algorithm is proposed in this paper. The method adds additional external force based on the fact that the two images can make the mutual information between them maximal, ,and the force is defined as the gradient of mutual information between the two images with respect to the deformation field. Moreover, to avoid local extrema and speed up the registration process, the algorithm is performed in a multi-resolution manner. Experiments are conducted with both mono-modality and multimodality images, the results show that this improved method can get a accurate registration transformation quickly.