An intermediate image generating means (1) generates a horizontal intermediate image (D1h) and a vertical intermediate image (D1v) by extracting components of an input image (DIN) in a particular frequency band; an intermediate image processing means (2) generates a horizontal image (D2Bh) and a vertical image (D2Bv) by performing non-linear processing (2A) and high-frequency component generation (2B); an intermediate image (D2) is obtained by combining these horizontal and vertical images by performing weighted addition for each pixel and is added (3) to the input image (DIN) to obtain an enhanced output image (DOUT). Even if the input image includes a fold-over component on the high-frequency side or does not include an adequate high-frequency component, adequate image enhancement processing can be carried out.