图像精准匹配技术的研究与仿真

Research and Simulation of Medical Image Anatomical Point Precise Matching Technology

研究图像的精准匹配问题。以医学解剖点的图象精确匹配技术为例,提出了一种基于高斯金字塔像素匹配算法的医学图像解剖点精准匹配技术。人体组织中的成分复杂,信息融合后,差异组织会形成像素灰度误差,传统的匹配算法偏重于空间映射,一旦像素点灰度信息存在误差,在进行空间映射过程中,会造成映射转换误差增大,后期配准位置点偏移。为了避免上述缺陷,利用尺度不变特征转换方法,提取医学图像中的解剖点特征。对图像解剖点特征进行误差补偿,最大程度修补位置偏差。利用高斯金字塔像素匹配方法,实现医学图像解剖点的匹配。实验结果表明,利用改进算法进行医学图像解剖点匹配,能够提高匹配的准确性,从而为手术提供准确的依据。

Research the precise matching of medical images anatomical points. The paper used scale-invariant feature transform method to extracted medical image anatomical point features. The errors of the characteristics of im- age anatomical point were compensated to Fix the position deviation. The Gaussian pyramid pixel matching method was used to realize medical image anatomical point matching. The experimental results show that the algorithm can improve the accuracy of the matching and can provide an accurate basis for operation.