医学图像的多方向加权层间插值算法

Weighted multidirectional interpolation algorithm for medical image

断层图像插值主要应用于医学图像的融合和三维重建,但传统插值方法通常是单方向的。文章提出一种在3层图像的多方向加权插值新算法,该插值利用逆梯度确定不同方向的权重,沿三维空间中物体边缘的方向进行计算,以达到较高的图像峰值信噪比。实验结果证实该算法能得到高精确度的插值图像。

The image slice interpolation is mainly applied to the fusion of medical image and its 3D reconstruction.The conventional interpolations are usually unidirectional.In this paper,a new weighted multidirectional interpolation over three levels of image slices is proposed.This kind of interpolation,with weights in different directions determined by the inverse gradient,is computed in the directions of boundaries of 3D objects so as to achieve high peak signal-to-noise ratio（PSNR）.Numerical experiments confirm that the proposed method can obtain high-precision interpolated images.