The image extracting apparatus includes the scene-change-image extracting unit (12) that extracts a scene change image located at a position at which a scene is changed from an image sequence acquired in time series by using a predetermined extraction condition, a comparing unit (132) that compares the scene change image with an adjacent image that is acquired at a time adjacent to the time at which the scene change image is acquired within a predetermined time-series range, and a summary image extracting unit (13) that extracts one of the scene change image and the adjacent image corresponding to the scene change image as a summary image by using the result compared by the comparing unit (132).