A document processing apparatus includes an image processing unit, a target determination unit, a text reading unit, and a keyword extraction unit. The image processing unit processes an image. The target determination unit determines whether an image that is a target of a process executed by the image processing unit is to be used as a keyword extraction target. The text reading unit reads a text from the image processed by the image processing unit. The keyword extraction unit extracts a keyword from the text read by the text reading unit from an image determined to be a keyword extraction target by the target determination unit.