A recognition apparatus includes: a memory configured to store a scene recognition dictionary relating a feature of a person included in distance information indicating a distance to the person with a scene identified by a formed angle between a plurality of reference axes perpendicular to each other and a vector based on a skeleton of the person and a skeleton recognition dictionary relating distance information corresponding to the scene with a joint position of the person; and a processor coupled to the memory and configured to: in a case where recognition distance information, which is distance information on an object person, is acquired, recognize a scene corresponding to the recognition distance information based on a feature of the recognition distance information and the scene recognition dictionary, and recognize a skeleton of the object person based on a recognition result of the scene recognized, the recognition distance information, and the skeleton recognition dictionary.