An image processing apparatus and method include providing a same label number to adjacent black pixels that form a group of black pixels, the label number provided to each group of black pixels being unique, the black pixels being included in binary read data, determining, when the number of black pixels in a group of black pixels counted on the basis of the unique label number is less than a predetermined number, that the black pixels with the unique label number are noise and removing the black pixels from targets to be processed, acquiring coordinate information on at least one of black pixels that remain after noise has been removed, and extracting a document area on the basis of the acquired coordinate information.