HDR Video Encoding and Decoding via Slice-Layer Reshaping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing HDR video encoding and decoding methods result in poor viewing quality due to irreversible information loss and high encoding overhead from dynamic range and color gamut conversions, particularly at high compression ratios.
Innovation Solution
Parse the video bitstream to obtain reshaping parameters from picture-layer and slice-layer data units, such as picture-layer parameter sets and slice-layer parameter sets, to dynamically adjust the reshaping process, reducing the need for frequent updates of other encoding parameters and minimizing encoding overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the reshaping parameter is adjusted for each picture to improve viewing quality, then the adaptive adjustment of encoding parameter is achieved, but the bit encoding overhead increases due to repeated encoding of other PPS parameters
Solution Approach 1:
The patent segments the parameter set structure by introducing a new slice-layer parameter set that contains only the reshaping parameter, separating it from the picture-layer PPS that contains other encoding parameters. This allows the reshaping parameter to be updated per picture without triggering updates of other parameters, thus reducing bit encoding overhead while maintaining adaptive adjustment capability.
Solution Approach 2:
The patent extracts the reshaping parameter from the picture-layer PPS and places it in a dedicated slice-layer parameter set. This extraction allows the reshaping parameter to be independently controlled and updated without affecting or requiring updates to other PPS parameters, solving the overhead problem while preserving adaptive adjustment benefits.
2Productivity
If the compression ratio is increased to reduce file size, then the encoding efficiency is improved, but the distortion due to lossy compression increases
Solution Approach 1:
The patent changes the reshaping parameter dynamically for each picture based on source characteristics and compression characteristics. By adjusting this parameter adaptively, the system optimizes the reshaping process for each specific picture, thereby reducing distortion even at high compression ratios while maintaining good encoding efficiency.
3Adaptability or versatility
If the dynamic range and color gamut conversions are performed to support HDR video, then the HDR video encoding is enabled, but the irreversible information loss occurs
Solution Approach 1:
The patent makes the reshaping parameter dynamic rather than static, allowing it to be adjusted for each picture based on source characteristics and compression characteristics. This dynamic adjustment optimizes the inverse reshaping process to minimize information loss while maintaining HDR video encoding capability.
Data Source
Figure 1~2
Figure 3~4
Figure 5~7
AI summary
The present disclosure provides picture encoding and decoding methods, picture encoding and decoding devices as well as a decoder and an encoder. The picture decoding method includes: parsing a video bitstream to obtain a reshaping parameter from a picture-layer and/or slice-layer data unit of the video bitstream, and determining a reshaping parameter used for reshaping a reconstructed picture according to the obtained reshaping parameter; and reshaping the reconstructed picture by using the reshaping parameter used for reshaping the reconstructed picture. The reconstructed picture is a picture obtained by decoding the video bitstream before performing the reshaping. The picture-layer and/or slice-layer data unit includes at least one of the following data units: a picture-layer parameter set and/or a slice-layer parameter set different from a picture parameter set (PPS), a parameter data unit which are included in an access unit (AU) corresponding to the reconstructed picture, slice header and a system-layer picture parameter data unit. With the present invention, the problems of poor viewing quality of a video and high encoding overhead of the reshaping parameter in the related art are solved.