An image processing apparatus capable of clipping a diagnostic region while retaining the pixel information of the original radiographic image even when the subject is imaged aslant. The image processing apparatus applies trimming processing to a radiographic image acquired by an imaging apparatus for acquiring a radiographic image of a subject and includes an image rotation unit for rotating the radiographic image so that the subject is aligned with a given direction to produce a display image, a diagnostic region designation unit for designating a first diagnostic region clipped with a diagnostic region designation frame represented by a rectangle for the display image, and a trimming processing unit for calculating a second diagnostic region in the radiographic image from a rotation amount of the display image and diagnostic region information representing the first diagnostic region and applying trimming processing to the radiographic image to clip the second diagnostic region.