Abstract: A weighted averaging algorithm of image zooming-out based on image subblock and corner points is presented. This algorithm overcomes the shortcoming of weighted averaging algorithm and leads to corner points distortion. First, according to the scaledown proportion the sub-block of every pixel at small image is obtained from original image, and then scans the block in corner-points image of the original image. If there is a corner point in the subblock, it can calculate the weighted map centered by this corner point. Otherwise, it calculates the weighted map centered by image center. This algorithm can keep more corner points than pure sub-block weighted algirithm by the method of weighted averaging for image zooming-out.

Key words: image zooming-out, weighted average, average value filter, corner detection