According to one embodiment, an image processing apparatus connected to a camera device that images a processing target includes: an image information acquisition unit; an accumulation subtraction image information creation unit; a feature amount information creation unit; and an action content identification unit. The image information acquisition unit sequentially acquires, from the camera device, image information formed by imaging the processing target thereby. Based on a temporal change of the image information acquired by the image information acquisition unit, the accumulation subtraction image information creation unit accumulates subtraction information for a predetermined period, which is made by motions of a person present in a room, and creates multivalued accumulation subtraction image information. The feature amount information creation unit creates feature amount information in the accumulation subtraction image information, which is created by the accumulation subtraction image information creation unit, from a region where there is a density gradient in the accumulation subtraction image information. The action content identification unit identifies an action content of the person present in the room from the feature amount information created by the feature amount information creation unit.