To make the conversation of a movie or the like and a text for which the conversation contents are written correspond by detecting voice and the corresponding one of the conversation and making them correspond based on the result of comparing conversation time written in the text with the time of the voice.
A moving image storage means 1 stores the moving images of the movie or the like, a voice recognition means 2 recognizes only voice information in the moving images, time per sequence is measured in a conversation time measurement means 3 and the result is held in a moving image time holding means 4. In the meantime, a conversation sentence storage means 5 stores the speeches or the like of a scenario. The information is changed to the voice in a voice generation means 6 and the time is measured in a voice time measurement means 1. A voice time holding means 8 holds the time, and for the result of comparing it with the time of the conversation held in the conversation time holding means 4 in a conversation time comparison means 9, the conversation in the moving image and the speech in the scenario are made to correspond in a moving image conversation and sentence correspondence means 10 and correspondence relation is held in a correspondence holding means 11.
Next Patent: ACTIVE IMAGE SENSOR PROVIDED WITH SHARED READ STRUCTURE