Many advanced video applications require manipulations of compressed video signals. Recently, the techniques such as frame-based motion compensation and scaling in the DCT domain for manipulating DCT coefficients have been proposed. However, these techniques are not properly applicable to MPEG2 video transcoding because the field-based motion compensation procedure is required in MPEG2 coding. In this paper, we propose a field-based motion compensation and a scaling technique in the DCT domain, and implement them for MPEG2 video transcoding. It has been proved that the proposed DCT domain techniques provide the same results as in the spatial domain. The techniques are successfully adopted to direct MPEG2 video transcoding in the DCT domain.
