An average value, minimum value, and maximum value data detecting unit detects an average value, a minimum value, and a maximum value in one frame of a red (R) signal. A difference value calculating unit calculates a difference value by subtracting the average value from the R signal. A processing unit estimates the average value to be a level of a background image occupying a large area of a screen, and generates a signal indicating a minimum value (concavity) at the average value and a maximum value (convexity) at the maximum or minimum value as a compensated object signal. An adding unit adds the compensated object signal to a composed image of basic depth models to generate depth estimation data for generating an image in a different viewpoint from a non-3D image, the image in the different viewpoint and the non-3D image constituting the pseudo 3D image.