An image processing apparatus includes an acquisition unit configured to acquire a document image, a primary region segmentation unit configured to segment the acquired document image into a plurality of regions, a detection unit configured to detect a text region including an erroneous sentence from the regions segmented by the primary region segmentation unit, a secondary region segmentation unit configured to detect a second attribute region partly overlapped with an original sentence of the erroneous sentence and separate the detected region into the second attribute region and a part of the original sentence, and a combining unit configured to combine the part of the original sentence separated by the secondary region segmentation unit with the text region including the erroneous sentence.