A video frame is input to a spectrum analyzer, which transforms the frame from a spatial domain into a frequency domain. A resolution determiner analyzes the output of the spectrum analyzer, thereby selecting a resolution appropriate for the video frame. Then, a resolution converter converts a resolution of the video frame into the resolution selected by the resolution determiner. Thereafter, the video frame is subjected to a coding process by block divider, DCT transformer, quantizer and variable-length coder so as to be output as a coded bit stream. By coding the input video frame at a resolution corresponding to the high-frequency component in the frame, a video of higher quality can be obtained with the bit rate reduced.