基于相关性的医学图像插值算法

Medical Image Interpolation Algorithm Based on Correlation

鉴于医学图像层间插值是图像三维重建的关键环节,它直接影响人体组织器官三维重建结果的准确性;为克服以往医学图像插值方法低效率,边缘模糊的缺陷,提出了一种新的医学图像插值算法;该方法利用人体断层扫描数据中体素相关性和组织相关性,对人体组织进行分类,采用以不同的方式对不同类别的点进行插值,并对这些点进行校验;实验结果表明,该算法在插值图像的精度以及时间上比现有的同类插值算法有了很大的改善,图像边缘更加清晰,能有效地应用于三维重建。

Considering the interpolation of cross--sectional medical images is one of the key steps of 3D medical image reconstruction and it directly affects the accuracy of the results on the three--dimensional reconstruction of human tissues and organs. To overcome the shortcomings such as blurred border of images and lower efficiency caused by present interpolation methods, an intersliee interpolation in view of the relativity for medical images is presented in this paper. This algorithm which makes good use of voxel relativity and structure relativity can classify human tissue, and then the different methods are adopted to interpolate the different points, In addition, error checkout is introduced to check the mismatching points. And the experimental validation and analysis indicate that our proposed algorithm has less computational complexity and improves the quality of image, and the edge of these images will be clearer. The result can be used to 3D reconstruction effectively.