Plural predetermined examination parts of a patient are imaged sequentially and still images of the respective examination parts are thereby acquired by inserting an endoscope insertion unit having an imaging optical system in its tip portion into the body cavity of the patient. To this end, the number of still images taken is counted every time one of the predetermined examination parts. The counted number of still images taken is compared with a preset number of images which corresponds to the number of the plural predetermined examination parts. An alert is generated when non-coincidence is detected between the counted number of still images taken and the preset number. The operator is thus notified of occurrence of a failure to image.