An image processing apparatus is provided, that is configured to: extract a first pixel value corresponding to a first viewpoint that is one of a plurality of viewpoints to capture a subject image, at a target pixel position from image data having the first pixel value; extract second and third luminance values corresponding to second and third viewpoints that are different from the first viewpoint, at the target pixel position from luminance image data having the second and third luminance values; and calculate at least any of second and third pixel values of the second and third viewpoints such that a relational expression between the second or third pixel value and the first pixel value extracted by the pixel value extracting unit remains correlated with a relational expression defined by the second and third luminance values.