Accurate motion detection is performed by discriminating whether a sensor detecting an object motion is mounted on a human body or not, and processing is executed with respect to metadata based on the result. Sensor information according to the motion is input from the sensor, and a sensor mounting position is determined. A sensor mounting position detection unit calculates a ratio between a high-frequency component and a low-frequency component included in the sensor information, and discriminates whether the sensor is mounted on the human body or is mounted on other than the human body, on the basis of the calculated ratio. A metadata generating unit inputs user motion detection information obtained by executing a motion detection algorithm assuming a sensor mounting position coincident with a sensor mounting position detection result, and generates the shot image corresponding metadata.