Distributive vide frequency coding method based on self adaptive Hashenhege type vector quantization

Abstract

The method comprises: 1) generating 'harsh' bit-stream, and using 'harsh' bit-stream to generate the edge information frame of motion compensation; 2) making inner-frame Wyner-Ziv coding for the Wyner-Ziv frame in order to generate Wyner-Ziv bit-stream; using the generated edge information frame to decode the Wyner-Ziv bit-stream; both steps above all use a LDPCA (low-density parity check codes) encoder with feedback.

Description

technical field [0001] The invention belongs to the technical field of video coding and decoding, in particular to a distributed video coding method based on adaptive "Hash" and lattice vector quantization. technical background [0002] Distributed Video Coding (DVC) is a novel video coding framework that has just been developed in recent years. Compared with other mature compression technologies (such as MPEG series, H.26* series, etc.), it has "simple coding" , a certain degree of robustness, etc., are suitable for emerging multimedia communications that require "friendly uploading" (such as mobile photography, large-scale sensor networks, network video surveillance, etc.). It makes use of the Slepian-Wolf and Wyner-Ziv side information source coding theory in information theory in the 1970s, trying to realize a distributed source coding framework of "independent coding and joint decoding" of related sources, and integrates the huge amount of calculation into motion search...

AI Technical Summary

Problems solved by technology

[0010] There are two requirements for "Hash", one is that it can carry as much information about the current coded block as possible, and the other is that the bit rate of "Hash" should be as low as possible compared to the Wyner-Ziv code stream, which involves Two questions: one is how to improve the ability of "Hash" to express the original coding block, and the other is how to compress "Hash" to obtain the lowest bit rate. The first method is represented by the information of "undersampling and coarse quantization" in the pixel domain Encoding block, the "hash" of "undersampling and coarse quantization" is not compressed and sent directly; method two uses DCT high-frequency coefficients to represent encoding block information, and entropy coding and run-length coding are used for quantized DCT high-frequency coefficients respectively, but these The processing of "Hash" adopts a method similar to intra-frame coding, which is compressed in a way close to the entropy of "Hash" itself, and the code rate of "Hash" is fixed

However, there is no in-depth research on the performance of "Hash" information at present, and the code rate of "Hash" information is fixed, and there is no "Hash" with adaptive code rate.

Images