Patent US 5,544,263

A circuit calculates motion vectors for use in compressing a motion picture signal. The motion picture signal is divided into portions, each portion representing a picture. The portions include a current portion representing a current picture. The current portion is subdivided into blocks. Each block represents a rectangular portion of the current picture, and includes a luminance signal, a first color signal, and a second color signal. A prediction picture circuit derives a prediction picture from the motion picture signal. A motion vector generating circuit generates a luminance motion vector, a first color motion vector, and a second color motion vector. The respective motion vectors are generated for a block by performing block matching between the luminance signal, the first color signal, and the second color signal of the block and signal components representing a block of the prediction picture corresponding to the block. A flatness judgment circuit judges whether the luminance signal component, the first color signal component, or the second color signal component of a matching prediction block of the prediction picture is not flat and, when one of the signal components is judged to be not flat, identifies the signal component determined to be not flat as a flatness judgment result. Finally, a motion vector selecting circuit selects the luminance motion vector, the first color motion vector, or the second color motion vector as the motion vector for the current block in response to the judgment result from the flatness judgment circuit.