图像放缩的离散算法

从简单的二值图像出发提出灰度图像的离散放缩算法。该算法尽可能地保持了放缩后图像内物体边界的清晰，自然，而放缩倍数可为任意非整数倍。应用该算法，图像放大后层面清晰，色彩丰富，忠实地反映了原始图像的面貌；图像缩小后能较好地体现原始图像的主要特征。实例证明，此方法比常见的商品软件有更好的边界处理效果，且计算稳定，实现方便，将算法已成功地应用于开发的医学图像处理系统中。

A Discrete Approach to Image Zooming

A discrete algorithm for grayscale image resizing through binary image zooming in 3D space is presented. The method can help to keep the edges and borders in the zoomed image as clear and natural as possible, with arbitrary real positive scaling factors. When an image is enlarged, it can be still with vivid color, and express the appearance of original image faithfully. When the size of an image is reduced, the most important features of the image can also be preserved. Experiments show that the new method is effective, stable and convenient for image zooming, and it is also more advantageous in border and edge preserving compared to many common commercial software. We have applied the algorithm in our medical image processing system.