Piecewise Cross-Color Channel Predictor for HDR Bandwidth
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for extending image dynamic range in displays are inefficient, particularly in transmitting and decoding high dynamic range (HDR) content, as they often require sending both HDR and standard dynamic range (SDR) versions, which consumes high bandwidth and can result in prediction errors due to global tone mapping and color clipping.
Innovation Solution
The implementation of a piecewise cross-color channel prediction (PCCC) method, where color channels in SDR images are segmented and assigned specific predictors to generate a predicted VDR image, using minimum mean-square error optimization for second-order models, and transmitting prediction parameters as metadata to decode VDR images efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If both HDR and SDR versions are transmitted, then display quality is improved, but bandwidth consumption increases
Solution Approach 1:
The patent extracts only the essential HDR information (residual data and prediction parameters) rather than transmitting complete HDR content. The SDR base layer is transmitted separately, and only the difference information needed to reconstruct HDR is sent, significantly reducing bandwidth while maintaining HDR display quality.
Solution Approach 2:
The patent applies piecewise prediction where different prediction models are used for different regions of the image (shadows, midtones, highlights). This local approach optimizes the prediction accuracy for each tonal region while minimizing the amount of data that needs to be transmitted to achieve high display quality.
2Device complexity
If global tone mapping is used, then processing simplicity is improved, but prediction accuracy deteriorates due to color clipping
Solution Approach 1:
The patent segments the tone mapping process into multiple pieces, applying different prediction models for different tonal regions (shadows, midtones, highlights). This segmentation avoids the color clipping problems of global tone mapping while maintaining processing efficiency, as each segment can be handled independently with appropriate prediction parameters.
Solution Approach 2:
The patent applies different prediction strategies to different local regions of the image based on their tonal characteristics. By using local quality assessment and region-specific prediction models, the system achieves high prediction accuracy without the color clipping artifacts that result from uniform global tone mapping.
3Measurement precision
If piecewise cross-color channel prediction is implemented, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent segments the prediction process into manageable pieces, using separate prediction models for different color channels and tonal regions. This segmentation improves prediction accuracy by capturing local variations while controlling computational complexity through efficient organization of prediction parameters that can be transmitted as compact metadata.
Solution Approach 2:
The patent uses parameter-based prediction models where prediction accuracy is controlled by adjusting parameters (such as regression coefficients for different color channels and regions). These parameters can be optimized to achieve high prediction accuracy while maintaining computational efficiency, and the parameters themselves become the compact data structure that needs to be transmitted.
Data Source
AI summary
A sequence of visual dynamic range (VDR) images may be encoded using a standard dynamic range (SDR) base layer and one or more enhancement layers. A prediction image is generated by using piecewise cross-color channel prediction (PCCC), wherein a color channel in the SDR input may be segmented into two or more color channel segments and each segment is assigned its own cross-color channel predictor to derive a predicted output VDR image. PCCC prediction models may include first order, second order, or higher order parameters. Using a minimum mean-square error criterion, a closed form solution is presented for the prediction parameters for a second-order PCCC model. Algorithms for segmenting the color channels into multiple color channel segments are also presented.