An image processing method includes decoding encoded image data to generate a decoded image signal including a luma signal and a chroma signal. The method further includes detecting a condition that a reference field has a different parity from a current field and that the number of chroma pixels is vertically different from the number of luma pixels. The method further includes shifting a phase of the chroma signal in the reference field in a vertical plus direction according a value of a motion vector in a vertical direction under the detected condition so that the reference field will coincide in phase of the chroma signal with a current field. The method also includes performing a motion compensation of the decoded image signal using the reference field.