An image processing apparatus generates a plurality of parallactic images having a relationship that the magnitude of an amount of image shift between parallactic images increases with an increase of a defocus amount, generates, through plural selections of the generated parallactic images permitting a repeated same selection, a series of parallactic images in which the selected parallactic images are arranged in a selection order; repeatedly applies interpolation processing to each pair of adjacent parallactic images included in the series of parallactic images to generate a plurality of interpolation images; and arranges the interpolation images between the adjacent parallactic images to output the interpolated parallactic images as an output image of a moving image.