In the present invention, a reference frame selection unit selects a reference frame from among a sequence of frames showing the heart of a fetus. A candidate region group setting unit sets a candidate region group for each of the frames that constitute the sequence of frames. A correlation value calculation unit calculates, for each candidate region, a correlation value between the reference frame and all the other frames. Due to this, a plurality of correlation value waveforms corresponding to the plurality of candidate regions is generated. A stabilized waveform portion specification unit specifies a stabilized waveform portion for each correlation value waveform. A stabilized region specification unit specifies, from among the plurality of stabilized waveform portions, the stabilized waveform portion having the highest degree of stabilization (in other words, the candidate region having the highest degree of stabilization). A heart rate calculation unit calculates heartbeat information (heart rate, etc) for the fetus on the basis of the stabilized waveform portion having the highest degree of stabilization.