An image format conversion apparatus including a 3:2 pull-down detection part determining whether a current field out of plural fields continuous in time is an image of a 3:2 pull-down format produced by a 3:2 pull-down process; a scroll text/graphic data detection part determining whether scroll motions exist on the current field; a 3:2 pull-down compensation part calculating interpolation values for pixels of the current field to be interpolated by using a 3:2 pull-down interpolation method; an IPC part calculating interpolation values for pixels of the current field to be interpolated by using a second interpolation method; and an output selection part for selectively outputting either the interpolation values of the 3:2 pull-down compensation part or the interpolation values of the IPC part based on results of the determinations of the scroll text/graphic detection part and the 3:2 pull-down detection part.