SDR-HDR Image Conversion Using Adaptive 2D LUT Luminance Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for converting between Standard Dynamic Range (SDR) and High Dynamic Range (HDR) images are computationally complex and inefficient, particularly for hardware implementations like FPGAs, and lack dynamic adaptation to image content.
Innovation Solution
A method and device using a 2D Look-Up-Table (LUT) to transform luminance components, employing adaptable parameters and a general variable computed per image, which simplifies computations and dynamically adapts output luminance to each image, allowing efficient conversion between SDR and HDR.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex tone mapping computations are used to convert between SDR and HDR, then conversion accuracy is improved, but computational complexity increases making hardware implementation difficult
Solution Approach 1:
The conversion process is segmented into two distinct stages: first computing a general variable (L med) that characterizes the overall luminance properties of the input image, and then using this general variable to index a pre-computed 2D LUT for obtaining the transfer function. This segmentation separates the complex computation into a simple image-wide calculation followed by a table lookup operation, dramatically reducing computational complexity while preserving conversion accuracy.
Solution Approach 2:
The 2D LUT containing the transfer functions is pre-computed and stored before runtime. The complex relationship between input luminance, general variable, and output luminance is established in advance, allowing the runtime conversion to simply compute L med and perform a table lookup rather than performing complex calculations for each pixel.
2Adaptability or versatility
If fixed conversion parameters are used, then device complexity is reduced, but adaptability to different image content deteriorates
Solution Approach 1:
The conversion process dynamically adapts to each input image by computing L med, which captures the specific luminance characteristics of that image. This dynamic variable is then used to index the 2D LUT, selecting the appropriate transfer function tailored to the current image's luminance distribution, achieving high adaptability without requiring complex real-time parameter adjustment mechanisms.
Solution Approach 2:
The general variable L med serves as an intermediary between the input image characteristics and the pre-computed 2D LUT. It translates the continuous variation in image luminance properties into a discrete index that selects the most appropriate pre-computed transfer function, enabling adaptability through a simple intermediary calculation rather than complex real-time optimization.
3Measurement precision
If complex real-time computations are performed for each pixel, then conversion precision is improved, but processing speed deteriorates
Solution Approach 1:
The per-pixel computation is segmented and replaced: instead of performing complex tone mapping calculations for each pixel, the method computes a single image-wide L med value and then uses this to index a pre-computed 2D LUT. This segmentation transforms the computational burden from O(N) complex operations to O(1) simple operations plus a table lookup, dramatically improving processing speed while maintaining precision through the use of pre-computed accurate transfer functions.
Solution Approach 2:
The complex luminance transformation relationships are pre-computed and stored in the 2D LUT before runtime. This preliminary action moves the computationally intensive work to an offline setup phase, allowing runtime processing to simply perform lightweight indexing and table lookups, thereby achieving high processing speed without sacrificing the precision of complex tone mapping.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
The invention concerns a method for converting an input image comprising an input luminance component made of elements into an output image comprising an output luminance component made of elements, the respective ranges of the output luminance component values and input luminance component element values being of different range extension. the method comprises for the input image: computing a value of a general variable representative of at least two input luminance component element values; transforming each input luminance component element value into a corresponding output luminance component element value according to the computed general variable value; and converting the input image using the determined output luminance component element values. The transforming step uses a set of pre- determined output values organized into a 2D Look-Up-Table (2D LUT) comprising two input arrays indexing a set of chosen input luminance component values and a set of chosen general variable values respectively, each pre-determined output value matching a pair of values made of an indexed input luminance component value and an indexed general variable value, the input luminance component element value being transformed into the output luminance component element value using at least one predetermined output value.