Distributed video coding correlated noise model construction method based on multi-probability distribution

Abstract

The invention discloses a distributed video coding correlated noise model construction method based on multi-probability distribution. The method comprises the follow steps: performing DCT transformation on correlated noises to obtain coefficients of 64 sub-bands; calculating the probability distribution of each sub-band, and calculating corresponding entropy through the probability distribution; calculating probability densities of each sub-band in three distributions including the Cauchy distribution, the Laplace distribution and the Gaussian distribution, and calculating the entropy of each sub-band in the three distributions; and finally, comparing the entropy of each sub-band with the absolute value of the difference of the entropy of the sub-band in the three distributions including the Cauchy distribution, the Laplace distribution and the Gaussian distribution separately so as to select the probability distribution having a minimum absolute value to perform modeling on the sub-band. By adopting the method of the invention, the disadvantage that the distribution of each sub-band is fail to be accurately described with an appropriate probability distribution with existing correlated noise model construction methods can be overcome, so the advantage of higher modeling accuracy can be realized, so that the rate-distortion performance of a distributed coding and decoding system can be improved.

Description

technical field [0001] The invention relates to distributed video coding technology, in particular to a method for constructing a correlation noise model in distributed video coding. Background technique [0002] Compared with traditional video coding standards such as MPEG-X and H.26X, Distributed Video Coding (DVC) has successfully transferred the computational complexity of the encoding end to the decoding end. It has the characteristics of simple encoding and complex decoding, and is very suitable for mobile video Wireless terminals with limited computing power and power consumption at the coding end, such as telephones, wireless video surveillance, and wireless video sensor networks. In DVC, correlated noise is defined as the difference between the original Wyner-Ziv (WZ) frame and the side information (SI) frame. In order to effectively use the check bit information sent by the encoder to correct side information errors, the distributed video decoder needs to use ...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to overcome the shortcomings of the existing correlation noise model construction method that fails to accurately describe the distribution of each sub-band with a suitable probability distribution, resulting in low model accuracy, and provides a correlation noise model based on multiple probability distributions Construct method to obtain higher modeling accuracy, thereby improving the performance of distributed codec system

Images