A filtering unit applies a filter for suppressing a high-frequency component to a second image generated by performing image processing including at least enlargement processing on a first image captured by using an endoscope. An image reducing unit generates a third image by performing reduction processing on the second image to which the filter has been applied. A filter setting unit sets a characteristic of a filter based on information on a maximum frequency component that can be reproduced in at least the first image and information on a maximum frequency component that should be reproduced in the third image.