Rate distortion optimization rapid decision-making system and method based on deep learning in HEVC intra-frame coding

Abstract

The invention discloses a rate distortion optimization rapid decision-making system and method based on deep learning in HEVC intra-frame coding, and belongs to the technical field of video coding. According to the method, the internal relation existing during coding parameter selection is considered, and the CU / PU partition and the prediction mode of the current coding unit are determined at thesame time in combination with the spatial-temporal correlation of the video content, so that the traversal process in the rate distortion optimization process in HEVC intra-frame coding is avoided, the calculation complexity can be reduced to the maximum extent, and the coding time is saved. The method specifically comprises the following steps: 1) preparing a video data set for training and testing a decision network; 2) respectively training the single-step decision network under different QPs by utilizing the training data set, and storing trained network model parameters; 3) embedding thesingle-step decision network into an HEVC reference model HM.15. 0 to realize a low-complexity HEVC encoder, and 4) respectively using the single-step decision network models trained under different QPs to encode the video on the test set, and testing the encoding complexity and RD performance.

Description

technical field [0001] The invention belongs to the technical field of video coding, and in particular relates to a rate-distortion optimization fast decision-making system based on deep learning in HEVC intra-frame coding and a method thereof. Background technique [0002] Video coding technology is the foundation of video services. Due to the huge amount of raw video data directly obtained from visual sensors, in order to effectively transmit and store video, the raw video is compressed with high bit rate-distortion (RD) quality and low complexity. Coding is necessary. In recent years, the further improvement of video resolution and the real-time requirements of video services have put forward higher requirements for video coding technology. As the latest video coding standard, the high-efficiency video coding standard H.265 / HEVC can achieve a significantly high compression ratio. However, H.265 / HEVC introduces more encoding parameters, and it is necessary to determine t...

AI Technical Summary

Problems solved by technology

First, from the perspective of complexity reduction, the CU / PU partition decision and prediction mode selection are an overall traversal process, which together leads to a large computational complexity of the HEVC encoder.

Therefore, focusing on only one of the tasks does not minimize the complexity

Second, there is an obvious correlation between the CU / PU partition decision and the prediction mode selection, and it is unwise to simply treat these decisions as a separate process and make decisions in two stages

Images