Image Reconstruction Using Candidate Color Space Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing mapping of SL-HDR metadata variables in ETSI TS 103 433-1 v1.2.1 is not optimal, leading to potential differences in reconstructed color volumes and spaces, resulting in visual artifacts and computational intensity due to the limited values mapped to 'hdrDisplayColourSpace' (0, 1, or 2), which do not accurately represent the vast possible combinations of Mastering Display Colour Volume (MDCV) SEI message values.
Innovation Solution
A method is proposed to map the given color space to a candidate color space that encompasses all primaries of the given color space within its convex hull, ensuring no clipping during image reconstruction, and reconstruct the image using this candidate color space, which involves comparing color volumes using metrics like DeltaE and selecting the candidate with the minimum score for accurate color matching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If the existing mapping of SL-HDR metadata variables is used, then the implementation is simple, but the color reproduction accuracy deteriorates due to limited mapped values (0, 1, or 2) that do not accurately represent vast MDCV SEI message combinations
Solution Approach 1:
The patent changes the parameter representation by mapping MDCV SEI message values to a more granular set of candidate color space values. Instead of using limited discrete values (0, 1, or 2), the system uses multiple candidate values that better represent the continuous range of color volumes, thereby improving color reproduction accuracy while maintaining implementation feasibility through automated selection algorithms.
2Productivity
If the existing mapping is used, then the processing is faster, but visual artifacts increase due to clipping in reconstructed color volumes
Solution Approach 1:
The patent introduces a dynamic selection mechanism that automatically chooses the most appropriate candidate color space from multiple options based on the specific MDCV SEI message values. This dynamic approach ensures that the selected candidate color space always encompasses the given color space without clipping, thereby eliminating visual artifacts while maintaining processing efficiency through pre-computed candidate sets.
3Device complexity
If the existing mapping is used, then the device complexity is lower, but computational intensity increases due to the need to handle limited mapped values that require complex reconstruction algorithms
Solution Approach 1:
The patent performs preliminary action by pre-computing and storing multiple candidate color spaces that can potentially encompass a given color space. During runtime, the system only needs to compare the MDCV SEI message values against these pre-computed candidates and select the best match, significantly reducing computational intensity compared to performing complex color volume reconstruction algorithms from scratch during processing.
4Manufacturing precision
If a candidate color space that encompasses all primaries is selected, then color clipping is eliminated, but the device complexity increases due to the need to compare and select from multiple candidate color spaces
Solution Approach 1:
The patent uses copying by creating multiple candidate color spaces that are copies or variations of standard color spaces (such as BT.709, BT.2020, etc.), each with slightly different parameters. These candidate copies are stored in advance, and the system simply needs to copy the appropriate candidate and use it for reconstruction, avoiding the need for complex real-time color space transformations while ensuring accurate color reproduction.
Data Source
AI summary
At least one embodiment obtain a method for reconstructing an image from metadata representing a given color space, comprising mapping said given color space to a candidate color space which encompasses in its convex hull all the primaries of the given color space; and reconstructing the image using said candidate color space.