Real-time video compression based-on three-dimensional matrix DCT

This paper presents a new video compression method based on the three-dimensional matrix DCT (3D-MDCT). This method adequately considers the correlation between the intraframe and interframe while regards them as a whole and makes them into a single mathematics model. 3D-MDCT algorithm is used in video system to eliminate the need for motion vectors computations or motion compensation computations. Experiments based on the parallel DSP system show that the proposed method can compress video image in real-time and the compression performance is better than MC/2D-DCT used in MPEG.