The video input unit inputs video data. The reference data input unit inputs a plurality of reference data relating to the video data. The extraction condition determining unit determines, from the identification information of the plurality of reference data, an item of identification information in correspondence with a period of time elapsed after receipt of an output request. The reference data extracting unit extracts the reference data having the determined identification information. The analyzing unit calculates the level of similarity between the frames in the input video data and the reference data, and extracts frames whose level of similarity is greater than a first threshold value. The video generating unit generates output video data that includes the extracted frames. The output unit outputs the generated output video data.