HDR Image Encoding with Luminance-Weighted Color Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image coding technologies struggle to effectively transition between images with different luminance dynamic ranges while maintaining chromatic accuracy, particularly in the transition from High Dynamic Range (HDR) to Standard Dynamic Range (SDR) for video communication systems.
Innovation Solution
A system that employs luminance mapping functions to convert HDR images into lower dynamic range images, utilizing metadata to guide the decoding process and adapt to various display capabilities, ensuring accurate chromatic representation across different luminance ranges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If luminance mapping is applied to convert HDR images to SDR, then compatibility with legacy displays is improved, but chromatic accuracy deteriorates
Solution Approach 1:
The patent applies different processing strategies to different regions of the image based on luminance levels. Bright highlights use one mapping approach while dark shadows use another, with smooth transitions between them. This local differentiation allows preservation of chromatic accuracy in critical regions while achieving overall display compatibility.
Solution Approach 2:
The mapping function is designed to be dynamic rather than static, adapting its characteristics based on the local luminance context. The function adjusts its slope and curvature differently across the luminance range, allowing it to maintain chromatic fidelity where needed while achieving compression for display compatibility elsewhere.
2Device complexity
If simple luminance compression is used, then device complexity is reduced, but chromatic fidelity deteriorates
Solution Approach 1:
The luminance range is segmented into multiple regions (bright highlights, mid-tones, dark shadows), each processed with appropriately tailored mapping characteristics. This segmentation allows the system to maintain chromatic fidelity in critical regions without requiring complex processing across the entire luminance range, balancing simplicity and accuracy.
Solution Approach 2:
The patent introduces intermediate calculation steps and auxiliary functions that mediate between the simple compression requirement and the fidelity requirement. These intermediary elements add computational overhead but preserve chromatic accuracy while achieving the desired luminance compression.
3Productivity
If aggressive luminance compression is applied, then dynamic range reduction is improved, but color saturation deteriorates
Solution Approach 1:
The mapping function is constructed as a composite of multiple mathematical segments with different characteristics. Each segment is optimized for specific luminance ranges and color preservation requirements, combining to achieve both aggressive compression and color saturation preservation through their coordinated action.
Data Source
Figure 1A~1D
Figure 2
Figure 3A~3C
AI summary
To obtain better color control in dynamic range mappings, which are e.g. useful in particular for HDR image coding and communication, the inventor proposes, for a corresponding innovative decoder, an encoder for encoding an original image (Im_MAST) of pixels, wherein a pixel has a primary brightness, wherein the original image is represented as a different communication image (Im_COMM) having for the pixel a secondary brightness, wherein the secondary brightness lies within a secondary range having a lower maximum brightnesses (ML_C) than an original maximum brightness (ML_V) of a primary range of the primary brightness, wherein the encoder comprises a first circuit (600) arranged to calculate an initial approximation of the secondary brightness (Y'_Ek_SDR) for a pixel being processed, the first circuit comprising: a largest color component determining circuit (506) arranged to for the pixel determine which is the largest one of a red, green and blue color component representing a color of the pixel, and outputting such largest color component (LC_i); a weight determination circuit (508) arranged to obtain a value of a control parameter (Ksi) and to calculate a first weight (A) by taking the minimum of the constant 1.0 and a result of a multiplication of the control parameter by a first intermediate value which is equal to the value of a second intermediate value minus 1.0, wherein the second intermediate value equals a division of the largest color component by an input luma (Y'_i) of the pixel; wherein the weight determination circuit is arranged to output the first weight and a second weight which is equal to 1.0 minus the first weight; a first multiplier determination circuit (604) arranged to output a luma-dependent multiplier (gd(Y')) which depends on the value of the input luma; a second multiplier determination circuit (607) arranged to output a largest component-dependent multiplier (gd(LC)) which depends on the value of the largest color component (LC_i); a first multiplication circuit (510) arranged to obtain a weighted luma multiplier (gd_W) by multiplying the luma-dependent multiplier (gd(Y')) by the second weight (1-A); a second multiplication circuit (511) arranged to obtain a weighted largest component multiplier (gd_LCW) by multiplying the largest component-dependent multiplier (gd(LC)) by the first weight (A); an adder (512) arranged to add the weighted largest component multiplier (gd_LCW) to the weighted luma multiplier (gd_W) to obtain a final multiplier (gFe); and a scaling multiplication circuit (513) to obtain the initial approximation of the secondary brightness (Y'_Ek_SDR) by multiplying the final multiplier (gFe) by the input luma (Y'_i); wherein the image encoder comprises a second circuit (700) arranged to yield an improved accuracy second approximation of the secondary brightness (Y'_iko) on the basis of a difference (DEL) between the input luma and an estimate of the input luma derived by luminance up-mapping the initial approximation of the secondary brightness (Y'_Ek_SDR) with a reciprocal value of the final multiplier (gFe).