A video multi-reference frame encoding method

Abstract

The invention discloses a multi-reference frame coding method for video, comprising: determining the current frame type and a reference frame list, if the current frame is a P frame, setting the reference frame list as the forward list List0; if it is a B frame, setting the reference frame list The frame list is two lists of forward List0 and backward List1; obtain the motion vector difference of adjacent PUs of the current PU, and predict the motion vector difference MVDP of the current PU; according to the storage order of the reference frames in the reference frame list, select each A reference frame; under the selected reference frame, obtain the motion vector difference Dc of the current PU through motion estimation and motion compensation; according to MVDP and Dc, determine whether to terminate the selection of other reference frames in the current reference frame list. The multi-reference frame rapid selection algorithm of the present invention does not change the code stream structure, and the generated code stream is compatible with any standard decoding, so it is not necessary to modify the playback client of the decoded video, and the application is convenient.

Description

technical field [0001] The invention relates to a video multi-reference frame encoding method. Background technique [0002] In the past few decades, people have higher and higher requirements for video quality, which has promoted the rapid development of video coding technology, resulting in many excellent video coding standards. The emergence of each generation of video coding has made video coding more efficient. Get a qualitative improvement. In January 2013, the International Telecommunication Union (ITU) officially announced the launch of a new generation of video codec standard HEVC (High Efficiency Video Coding). HEVC provides a more flexible, reliable and stable video solution than H.264. For video files of the same quality, the file size encoded using the HEVC standard is half that of the H.264 standard, but the technology and algorithm it uses are more than the previous ones. Standards are more complex and require more computation during video compression. [0...

AI Technical Summary

Problems solved by technology

However, since a large number of complex and high rate-distortion cost calculations are involved in the motion search process, the multi-reference frame technology makes the process require the same processing for all reference frames, which dramatically increases the computational complexity of the reference frame selection process.

Images