Cascaded Chroma Reshaping for HDR Video Signal Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression standards struggle with efficient chroma reshaping of high-dynamic range (HDR) images, particularly in the ICtCp color format, where chroma errors are magnified and not adequately addressed by current methods that only consider luminance values, leading to visible distortions, especially in low and high luminance ranges.
Innovation Solution
The proposed solution involves a cascaded chroma reshaping (CCR) method that uses both luminance-based (LCR) and saturation-based (SCR) scaling functions to reshape chroma components, with LCR and SCR scalers being computed and applied to generate reshaped chroma components, and inverse reshaping processes to reconstruct original signals, considering the ICtCp color format's unique characteristics and human visual system response.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If chroma reshaping is applied using only luminance-based scaling, then coding efficiency is improved, but chroma artifacts become visible especially in low and high luminance ranges
Solution Approach 1:
The chroma reshaping process is segmented into two independent stages: luminance-based chroma reshaping (LCR) and saturation-based chroma reshaping (SCR). Each stage addresses specific aspects of chroma distortion, with LCR handling luminance-dependent artifacts and SCR handling saturation-dependent artifacts. This segmentation allows each component to be optimized independently while working together to resolve the overall chroma artifact problem.
Solution Approach 2:
The invention adds a new dimension to chroma reshaping by introducing saturation-based scaling in addition to the existing luminance-based scaling. Instead of relying solely on luminance values (one dimension), the system now operates in a two-dimensional space considering both luminance and saturation. This dimensional expansion enables the system to address chroma artifacts that were previously invisible or unaddressable in luminance-only approaches.
2Object-affected harmful factors
If chroma reshaping is applied using both luminance-based and saturation-based scaling, then chroma artifacts are reduced, but computational complexity increases
Solution Approach 1:
By segmenting the chroma reshaping into separate LCR and SCR stages, each stage can be implemented with dedicated, optimized computation. The LCR stage processes luminance-based scaling independently, and the SCR stage processes saturation-based scaling independently. This segmentation allows for more efficient implementation compared to a single unified approach, as each stage can leverage specific properties of its input data.
Solution Approach 2:
The system applies partial reshaping actions selectively. Not all chroma components require full reshaping processing - the LCR and SCR stages can be applied with different levels of intensity depending on the specific content characteristics and artifact severity. This allows the system to achieve adequate artifact reduction with reduced computational effort in cases where full processing is not necessary.
Data Source
Figure 1A~1B
Figure 2
Figure 3A~3B
AI summary
Methods for the chroma reshaping of high-dynamic range (HDR) signals are presented. For each input pixel, a first scaler is generated based on the pixel luminance and a luma-based chroma reshaping (LCR) function.A second scaler is generated based on a saturation-based chroma reshaping (SCR) function and the saturation value of the pixel.A cascaded scaler is generated by multiplying the first scaler with the second scaler. Reshaped chroma values are generated by multiplying the input chroma values with the cascaded scaler. The characteristics of preferred LCR and SCR functions are discussed. Inverse decoding methods based on chroma-reshaping metadata that define the LCR and SCR functions are also described.