Abstract: A self-adaptive amending algorithm of image zooming-out based on weighted averaging image sub-block is presented in this paper.This algorithm effectively retains the integrity of gray image information.First，low-resolution image obtained by weighted averaging for image zooming-out is disposed by the way of contrast extension .Then the algorithm takes account of the influence on the whole image information by local area variance and self-adaptively amends the every pixel value of low-resolution image.Experimental results demonstrate that the algorithm can more effectively avoid blurring and artifacts edges than only by the method of weighted averaging for image zooming-out，and generate high quality images.Especially，the visual quality of the images for the large ratio of image zooming-out is more prominent.

Key words: image zooming-out, self-adaptive, local area variance, weighted average, average value filter

摘要： 提出了一种基于图像子块加权缩小的自适应修正算法，有效地保持了灰度图像信息的完整性。首先，通过对加权平均缩小后得到的低分辨率图像进行对比拉伸处理，然后根据局部区域方差对整体图像的影响情况，对缩小图像像素点进行自适应修正。实验证明，算法克服了加权平均缩小后图像的边缘锯齿和模糊效应，使修正优化后的图像清晰。尤其对缩小比率大的图像质量的提高有较显著的效果。

关键词: 图像缩小, 自适应, 局部区域方差, 加权平均, 均值滤波器