A character area is extracted from an original image, which is a digital image, according to a feature value calculated with respect to the original image. Pixels belonging to the extracted character area are classified into a first class and a second class according to colors, in each of blocks divided from the original image. A black character color and a ground color on the original image are estimated according to the classification of the pixels belonging to the extracted character area into the first and second classes. A tone correction is performed to the original image according to the estimated black character color and the estimated ground color so as to adjust a contrast of the black character and the ground in an output image.