An image processing apparatus includes a deleted differential image generator and a region specifying unit. The deleted differential image generator generates, from first image data generated by reading an image of an unfilled or filled-in sheet and second image data generated by reading an image of a sheet being in a format identical to the unfilled or filled-in sheet and having a superimposed sheet superimposed thereon, deleted differential image data representing an image that is not represented in the second image data but is represented in the first image data. The region specifying unit specifies a region where the superimposed sheet is represented in the second image data, based on a region where the deleted differential image data is.