Determining Optimal Cube for 3D-DCT Based Video Compression for Different Motion Levels

This paper proposes new three dimensional discrete cosine transform (3D-DCT) based video compression algorithm that will select the optimal cube size based on the motion content of the video sequence. It is determined by finding normalized pixel difference (NPD) values, and by categorizing the cubes as "low" or "high" motion cube suitable cube size of dimension either [16×16×8] or[8×8×8] is chosen instead of fixed cube algorithm. To evaluate the performance of the proposed algorithm test sequence with different motion levels are chosen. By doing rate vs. distortion analysis the level of compression that can be achieved and the quality of reconstructed video sequence are determined and compared against fixed cube size algorithm. Peak signal to noise ratio (PSNR) is taken to measure the video quality. Experimental result shows that varying the cube size with reference to the motion content of video frames gives better performance in terms of compression ratio and video quality.