Video Reshaping Using Piece-Wise Linear Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing reshaping mechanism for video coding in HEVC requires modifications to the codec and specification, limiting its compatibility and efficiency in managing signal ranges between input and output videos.
Innovation Solution
The proposed method utilizes the Colour Remapping Information (CRI) SEI message to manage the range of input and output videos through piece-wise linear functions, encoding these functions into a bitstream, and applying inverse reshaping based on targeted ranges, without modifying the HEVC encoder or decoder.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the existing reshaping mechanism is used to improve compression efficiency, then compression efficiency is improved, but modifications to the HEVC codec and specification are required
Solution Approach 1:
The patent introduces an intermediary conversion process where the input video is first transformed into a reshaped video using piece-wise linear functions before being encoded by the standard HEVC encoder. This intermediary step allows the system to achieve improved compression efficiency while keeping the core HEVC encoder unchanged, as the reshaping prepares the data in a form that is more amenable to standard HEVC compression algorithms.
Solution Approach 2:
The patent changes the parameters of the video signal by applying piece-wise linear transformation functions to reshape the input video. These parameter changes in the video data (through reshaping) enable better compression performance without requiring changes to the HEVC codec parameters or specification, as the transformation occurs in the video data domain rather than the codec domain.
2Productivity
If the reshaping mechanism is integrated into HEVC, then compression efficiency is improved, but compatibility with existing HEVC implementations is reduced
Solution Approach 1:
The patent segments the video encoding process into two independent parts: a reshaping module that transforms the input video using piece-wise linear functions, and the standard HEVC encoder that processes the reshaped video. This segmentation allows existing HEVC implementations to remain unchanged and compatible, while the reshaping module can be added as a separate preprocessing step to achieve improved compression efficiency.
Solution Approach 2:
The reshaped video acts as an intermediary between the input video and the HEVC encoder. By introducing this intermediate representation, the system maintains compatibility with standard HEVC decoders (which can decode the bitstream as usual) while still achieving improved compression efficiency through the preprocessing transformation.
3Manufacturing precision
If piece-wise linear functions are used for reshaping, then signal range management is improved, but additional side information encoding is required
Solution Approach 1:
The patent applies piece-wise linear functions with a limited number of pieces (typically 2-4 pieces) to achieve adequate signal range management. This partial action approach provides sufficient precision for most video content without requiring the full complexity of many-piece functions, thereby limiting the amount of side information that needs to be encoded while still achieving good signal range management.
Solution Approach 2:
The patent encodes parameters representing the piece-wise linear functions (such as breakpoints and slopes) as side information in the bitstream. By efficiently encoding these parameters using compact representations and leveraging the statistical properties of video signals, the system achieves precise signal range management while minimizing the additional bitstream overhead.
Data Source
Figure 1~3
Figure 4~7
Figure 8
AI summary
The present principles relates to a method and device for encoding an input video comprising a luma component and two chroma components. The method comprises: - reshaping of said input video by applying (610) piece-wise linear functions to the luma and chroma components; - encoding (620) the reshaped video into a bitstream; and - encoding (630) parameters representative of said piece-wise linear functions as side information; and the method is characterized in that reshaping of said input video further depends on a targeted range of a decoded reshaped video and the range of a reconstructed video after an inverse-reshaping is applied on the decoded reshaped video.