An image processing apparatus includes a command-data storage unit, a handwritten-data recognizing unit, and a matching unit. The command-data storage unit stores therein a command-data table that contains a command character and content of a command corresponding to the command character in an associated manner. The handwritten-data recognizing unit performs character recognition and image analysis on image data to extract handwritten information including a command graphic representing a command with respect to the image data and a command character handwritten near the command graphic. The matching unit matches the command character extracted by the handwritten-data recognizing unit with the command character in the command-data table.