拼接算法在磁共振脊柱图像处理中的应用

Application of Stitching Algorithm in MRI Spinal Image Processing

图像拼接算法已被广泛应用于医疗图像处理,其通过有效拼接多张图像来解决单张图像信息显示不全的问题。拼接流程分为图像预处理、配准、融合、拼接4个部分。考虑到配准算法是图像拼接流程的重点和难点,针对基于特征的配准算法及基于灰度信息的配准算法分别进行了分析和仿真,并结合实际临床需求,设计并开发完成了一套全脊柱拼接软件。该软件既能够实现全脊柱自动、手动拼接流程,又具有集成多种图像特征计算及矩阵变换算法的功能,为相关专业人员提供了算法研究及需求讨论的新途径。

algorithm of image stitching has been widely used in medical image processing,which can solve the problem of incomplete display of single image information by effectively stitching multiple images.The stitching process is divided into four parts:Image preprocessing,registration,fusion and stitching.Considering that registration algorithm is the focus and difficulty of image stitching process,this paper analyzes and simulates feature-based registration algorithm and gray information based registration algorithm respectively,designs and develops a set of whole spine stitching software combined with actual clinical requirements.The software not only can realize the automatic and manual splicing process of the whole spine,but also has the function of integrating a variety of image feature calculation and matrix transformation algorithms,which provides a new way for the algorithm research and demand discussion for related professionals.