Title :
A method for estimating coding gain of subband filter considering higher order statistic
Author :
Yokota, Yasunari ; Usui, Shiro
Author_Institution :
Fac. of Eng., Gifu Univ., Japan
Abstract :
In transform, or subband coding, a coding gain has been widely used for predicting coding efficiency and designing a subband filter. It is a useful criterion which represents performance of the transform, or the subband filter, for object coding. First of all, this paper points out that the estimated coding gain includes error when the coding object has non-normality, such as a signal operated by non-linear transform and actual image data. A previous method for estimating the coding gain has been constructed with an assumption that signal value of the coding object is normally distributed. We propose a new method for estimating coding gain more accurately from higher order statistics of the coding object and the subband filter coefficients. It makes use of a property that the probability density function of the subband series is considerably approximated as a generalized Gaussian distribution. Using Gaussian series, squared and cubed Gaussian series, and series cut from the standard image “Lena” as coding objects, and orthogonal wavelet filter as the subband filter; the coding gains are estimated by the proposed method and a previous method, and compared to the experimental coding gains. It is shown that the proposed method is more accurate than the previous method
Keywords :
Gaussian distribution; filtering theory; higher order statistics; image coding; normal distribution; wavelet transforms; coding efficiency; coding gain; cubed Gaussian series; generalized Gaussian distribution; higher order statistic; nonlinear transform; orthogonal wavelet filter; probability density function; squared Gaussian series; subband coding; subband filter; Discrete transforms; Discrete wavelet transforms; Entropy; Gaussian distribution; Higher order statistics; Image coding; Information filtering; Information filters; Nonlinear filters; Probability density function;
Conference_Titel :
Neural Networks for Signal Processing [1996] VI. Proceedings of the 1996 IEEE Signal Processing Society Workshop
Conference_Location :
Kyoto
Print_ISBN :
0-7803-3550-3
DOI :
10.1109/NNSP.1996.548370