An endoscope apparatus 1 includes a detecting section 34 to which an observation image G1 of a subject is sequentially inputted, the detecting section 34 being configured to detect a characteristic region L in the observation image G1 based on a predetermined feature value concerning the observation image G1, and an emphasis processing section 36a configured to apply, when the characteristic region L is continuously detected in the detecting section 34, emphasis processing of a position corresponding to the characteristic region L to the observation image G1 of the subject inputted after elapse of a first predetermined time period from a time when the characteristic region L is detected.