An image obtaining unit obtains actual endoscope images, and a virtual endoscope image generating unit generates virtual endoscope images including a plurality of virtual endoscope branch images. A corresponding virtual endoscope image determining unit obtains a plurality of actual endoscope images which were obtained within a predetermined amount of time before the endoscope reached its current position, compares the plurality of actual endoscope images and the plurality of virtual endoscope branch images, and determines a corresponding virtual endoscope image that corresponds to the branch structure closest to the current position of the endoscope, through which the endoscope has passed. A matching unit performs matching between each of a plurality of actual endoscope path images and a plurality of virtual endoscope path images for each of a plurality of paths. A position identifying unit identifies the current position of a leading end of the endoscope based on the results of matching.