A Novel Statistical Distortion Model Based on Mixed Laplacian and Uniform Distribution of Mpeg-4 FGS.

Xie Li,Wenjun Zhang

2006-01-01

Abstract:The Laplacian Distribution has been widely used to characterize the distribution of DCT coefficients in a great deal of research on statistical attributes of DCT for the reason that Laplacian distribution can achieve a good balance between the simplification of the model and fidelity to the real empirical data. But in our research, we found that there is a large discrepancy between the real data and the distortion model based on Laplacian distribution. Furthermore, the higher the rate, the larger the distortion. In this paper, we proposed a new distortion model based on mixed Laplacian and Uniform distributions. Experimental results show that the MLQ distortion model can approximate the real empirical data and has a higher degree of accuracy than the conventional distortion model based on Laplacian distribution. Compared with the conventional distortion model based on Laplacian distribution, the new model can decrease the MSE distortion by 2–6 units. The MLQ distortion model can represent the relationship of rate and distortion of the Mpeg-4 FGS enhancement layer with a higher degree of accuracy.