HDR Image Reshaping Maps for Reversible SDR Editing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies struggle to efficiently convert between standard dynamic range (SDR) and high dynamic range (HDR) images while maintaining color accuracy and compatibility with various display devices, leading to issues with reversibility and adverse editing impacts.
Innovation Solution
Implementing tensor-product B-Spline (TPB) based image reshaping solutions with optimized forward and backward mapping processes to generate SDR images from HDR images, using static or dynamic frameworks to ensure accurate reconstruction and compatibility across different devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing image processing technologies are used to convert between SDR and HDR images, then processing speed is maintained, but color accuracy and reversibility deteriorate
Solution Approach 1:
The patent pre-calculates and stores forward and backward mapping tables during an offline training phase. These pre-computed mappings are then directly applied during runtime conversion between SDR and HDR images, eliminating the need for complex real-time calculations and achieving both high color accuracy and fast processing speed.
Solution Approach 2:
The patent implements a dynamic framework that can adaptively select between different mapping strategies (forward mapping, backward mapping, or hybrid approaches) based on the specific image content and requirements. This dynamic selection optimizes both color accuracy and processing efficiency for different scenarios.
2Measurement precision
If complex mapping processes are used to maintain color accuracy, then color precision is improved, but device complexity increases
Solution Approach 1:
The complex mapping relationships are pre-computed during an offline training phase and stored in lookup tables. During runtime, the system simply performs table lookups and interpolations, which are computationally simple operations. This separates the complexity of finding optimal mappings from the complexity of executing the conversion.
Solution Approach 2:
The patent introduces intermediate representation spaces and uses them as mediators in the conversion process. By transforming images through carefully designed intermediate spaces with known mapping properties, the system achieves accurate color conversion while keeping the actual processing steps simple and well-defined.
3Adaptability or versatility
If HDR images are converted to SDR images for compatibility, then display compatibility is improved, but information loss increases
Solution Approach 1:
The system pre-calculates both forward (HDR to SDR) and backward (SDR to HDR) mapping tables during training. This allows the system to not only convert HDR to SDR for compatibility but also to reconstruct HDR images from SDR images, minimizing information loss by using optimized reconstruction algorithms that were pre-trained on large datasets.
Solution Approach 2:
The patent employs advanced parameter optimization techniques during the training phase to find mapping functions that minimize information loss. By adjusting mapping parameters and using techniques like perceptual uniformity optimization, the system preserves as much dynamic range information as possible during SDR conversion while maintaining display compatibility.
4Measurement precision
If advanced reshaping solutions are implemented, then reconstruction accuracy is improved, but power consumption increases
Solution Approach 1:
The computationally intensive parts of the reshaping process are performed offline during training to generate lookup tables and mapping functions. During actual image processing, the system uses these pre-computed resources requiring minimal computation, thus achieving high reconstruction accuracy with low power consumption on mobile devices.
Solution Approach 2:
The system uses pre-trained models and lookup tables that were optimized during development to perform efficiently on specific hardware platforms. The mapping processes are designed to leverage hardware acceleration and memory structures that minimize power consumption while maintaining high accuracy.
Data Source
AI summary
HDR color patches are sampled throughout an HDR color space parameterized by a parameter. Reference SDR color patches, input HDR color patches and reference HDR color patches are generated from the sampled HDR color patches. An optimization algorithm is executed to generate an optimized forward reshaping mapping and an optimized backward reshaping mapping. The optimized forward reshaping mapping is used to forward reshape input HDR images into forward reshaped SDR images, whereas the optimized backward reshaping mapping is used to backward reshape the forward reshaped SDR images into backward reshaped HDR images.