According to one embodiment, an image processing apparatus includes an image-capturing module configured to captures an image, a module configured to detect a first object region from the image, a module configured to extract color information of an image of the first object region, a module configured to detect candidates of a second object region, which is an object of recognition, from the image, a module configured to extract color information of an image of the second object region, and sets reference color data, a module configured to select the candidates of the second object region on the basis of the reference color data and the color information of the image of the first object region with respect to each of the candidates of the second object region, and a module configured to output, as an object of recognition, any one of the candidates of the second object region.