| 摘要 |
<p>In an image processing method, a square submatrix formed from adjacent original pixels is equally divided into small square regions along the X- and Y-coordinate axes. Approximate points are set at the vertexes of the regions. Interpolation coefficients are derived based on a predetermined interpolation function. Interpolation coefficients normalized so as to adjust the sum of the coefficient values of interpolation coefficients used for one interpolation operation to 2<k> (k is a positive integer) are calculated, and stored in a coefficient buffer in advance. A new pixel position of each pixel constituting a new image is calculated in accordance with magnifications representing enlargement/reduction ratios along the X- and Y-coordinate axes for an original image. An approximate point closest to the new pixel position is selected as the approximate point of the new pixel position. Interpolation coefficients corresponding to the interpolation original pixels are read out from the coefficient buffer. Interpolation operation is performed by product-sum operation, thereby calculating a pixel value at the approximate point. The calculated pixel value is divided by 2<k> to output a pixel value at the new pixel position. An image processing apparatus is also disclosed. <IMAGE></p> |
