A Video Coding Hard Decision Quantization Method Based on Content Adaptive Offset Model

Abstract

The invention discloses a video coding dead zone quantization method based on content self-adaptation, which adopts a content self-adaptive quantization offset model to realize a content self-adaptive quantization algorithm. This model can realize adaptive change of quantization offset according to different video contents. At the same time, in terms of rate-distortion performance, compared with the dead-zone HDQ algorithm with fixed offset, it has significantly improved; because the algorithm uses independent hard-decision quantization of coefficients, quantization can be processed in parallel at the coefficient level, which is easy to implement on hardware. The invention obtains relatively accurate distribution parameters based on the independent statistical analysis of DCT coefficients, and constructs a coefficient-adaptive quantization offset model by using the distribution parameters and quantization parameters. In video coding and quantization, the adaptive dead-zone HDQ quantization based on this model is adopted to realize high-efficiency video coding.

Description

technical field [0001] The algorithm of the present invention is applicable to the quantizer design of H.264 video encoder, and is also applicable to the quantizer design of H.265 / HEVC video encoder, specifically a video encoding hard-decision quantization method based on content adaptive offset model . Background technique [0002] Quantization is the core module in lossy video compression encoders such as MPEG and H.26x. Quantization maps the transformation coefficients with a large dynamic range to a limited number of quantization intervals, making full use of the perceptual redundancy characteristics of the human visual distortion system. Signal compression with little perceived distortion. Quantization determines the size of video compression distortion, and also has a great impact on bit rate control. The video codec standard stipulates the quantization step size parameter (Qp), and specifies the details of the proposed quantization, but does not make specific provis...

AI Technical Summary

Problems solved by technology

[0014] The defect of this prior art: this algorithm is the simplification to SDQ algorithm, but is mainly aimed at video encoder based on software implementation, and fails to eliminate the seriality caused by Viterbi algorithm branch selection and context arithmetic coding

Therefore, the extremely high correlation between the data still hinders the implementation of the algorithm on hardware

Images