An image processing apparatus and method that can improve encoding efficiency while preventing an increase in load. An extraction circuit of a filtering prediction circuit extracts motion compensation images for generating a prediction image in a high-resolution enhancement layer from reference frames in a low-resolution base layer. A filtering circuit of the filtering prediction circuit performs filtering, which involves upconversion and which uses analysis in the time direction, on a plurality of motion compensation images in the base layer extracted by the extraction circuit to generate a prediction image in the enhancement layer. The image processing apparatus may be an encoding apparatus or a decoding apparatus.