To speedily convert an image that is compressed and coded in the directions of spatial axis and time axis to an image that is compressed and coded and has different resolution.

A variable length decoding means 101 demodulates the compression image and outputs the movement vector information of each frame and an orthogonal transformation image. The orthogonal transformation image is subjected to inverse quantization, a hierarchy inverse orthogonal transformation means 103 performs the inverse orthogonal transformation of an orthogonal transformation image according to the change rate of the number of pixels, and outputs a non-compression differential image or a non-compression image after resolution change. A motion vector correction means 104 corrects decoded motion vector information according to the change rate of the number of pixels. A motion vector generation means 105 uses decoded motion vector information to generate the motion vector information after the resolution change, uses the motion vector information to allow a compensation non- compression image generation means 106 to compensate for the motion of the non-compression differential image to generation the non-compression image. A motion compensation non- compression difference image generation means 107 uses the motion vector information after the resolution change to generate the non-compression differential image of the non- compression image. The non-compression differential image and the non-compression image are compressed next. The non-compression image after the resolution change can be generated directly from the compression image before the resolution change.