An image area most similar to the image area of the image input signal is searched in the reference image from the frame memory 10 to generate a motion vector. A motion-compensated reference picture is generated from the motion vector and the reference picture from the frame memory. A prediction residual is generated by subtracting the motion-compensated reference image from the video input signal. The reference image stored in the memory is generated by adding the result of the orthogonal transform processing of the prediction residual, the quantization processing, the inverse quantization processing, and the inverse orthogonal transform processing to the motion compensated reference image. And the coded video output signal is generated by the orthogonal transformation processing of the prediction residual, the quantization processing, and the variable length coding processing. The reference image includes the in-screen reference images A, B and C inside the video display screen and the out-of-screen reference image D outside the video display screen, and the out-of-screen reference image D includes the in-Are generated based on the positional relationship between a plurality of similar reference images A and B. [;
展开▼
