An approach to adaptive 3D-DCT based motion level prediction algorithm for improved 3D-DCT video coding

摘要

This paper reports an adaptive three dimensional discrete cosine transform (3D-DCT) based motion level prediction algorithm which determines the optimal cube for 3D-DCT based compression technique by analyzing the motion content of the video sequence. Irrespective of the motion levels in the video sequence the generally used cube size is [8 × 8 × 8], but the proposed algorithm reported in this paper will adaptively choose the cube size in relation to the motion level of video sequence. The effectiveness of the algorithm can be verified by performing Rate Vs distortion comparison with different motion level sequences. Peak Signal to Noise Ratio (PSNR) has been taken as a measure of distortion. Experimental results reveals that without any motion compensation technique ,the proposed 3D-DCT algorithm which adaptively selects the cube size relative to the motion content of video sequence gives better performance in terms of reduction in the data rate and speed up the encoding process compared to the existing 3D-DCT based video compression algorithm.%W artykule opisano algorytm kompresji video bazujący na adaptacyjnej trójwymiarowej dyskretnej transformacie kosinusowej i przewidywaniu poziomu ruchu. W sposób adaptacyjny dobierany jest rozmiar sześcianu. Weryfikacji dokonano przez porównanie szybkości względem zniekształceń dla różnych sekwencji ruchu.