Image Tone Conversion Using Variable Luminance Step Size
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image tone conversion techniques are not optimized for specific image devices, leading to inefficient bit allocation and potential visual tone gaps, especially when dealing with wider dynamic ranges.
Innovation Solution
An image processing apparatus that determines a tone-conversion characteristic by calculating a maximum luminance difference threshold based on human visual perception models, such as JND, to map input image levels to output levels efficiently, considering both input and output device characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If non-linear quantization is performed considering human luminance discrimination capabilities (JND), then visual tone gaps are suppressed, but the technique is not optimized to specific image devices and bit allocation efficiency is reduced
Solution Approach 1:
The patent applies local quality by differentiating tone conversion processing between dark and bright luminance regions. In dark regions, fine quantization is applied to suppress visual tone gaps, while in bright regions, coarse quantization is used to reduce data amount. This localized approach optimizes each region according to human visual perception characteristics and device capabilities.
Solution Approach 2:
The patent changes the quantization parameter (bit depth) based on luminance level. The decision unit determines a maximum step size d such that the luminance difference between adjacent quantized levels remains within a just-noticeable difference threshold. This dynamic parameter adjustment optimizes both visual quality and data efficiency for specific image devices.
2Object-affected harmful factors
If fine quantization is applied in dark luminance regions, then visual tone gaps are suppressed, but data amount increases
Solution Approach 1:
The patent applies different quantization strategies to different luminance regions. Fine quantization with smaller step sizes is applied only in dark regions where visual tone gaps are more noticeable, while coarse quantization is applied in bright regions, thereby suppressing visual artifacts while minimizing overall data amount.
Solution Approach 2:
The quantization step size parameter is dynamically changed based on the luminance level of each pixel. The decision unit calculates the maximum allowable step size d for each region, allowing fine quantization where necessary and coarse quantization where acceptable, thus balancing visual quality and data efficiency.
3Quantity of substance
If coarse quantization is applied in bright luminance regions, then data amount is reduced, but visual tone gaps may arise
Solution Approach 1:
The patent dynamically adjusts the quantization step size parameter based on luminance level and device characteristics. In bright regions, the decision unit determines an appropriate step size d that allows coarser quantization while maintaining luminance differences within just-noticeable difference thresholds, thus reducing data amount without creating visible tone gaps.
4Measurement precision
If the number of bits is increased to record wider dynamic range images, then luminance precision is improved, but data amount and transmission cost increase
Solution Approach 1:
The patent changes the bit depth parameter dynamically based on luminance level rather than using a fixed high bit depth for the entire image. The tone conversion processing with variable step size d preserves luminance precision in dark regions where it is most critical, while using fewer bits in bright regions, thus maintaining overall luminance precision while reducing total data amount.
Solution Approach 2:
The patent applies different bit depths to different luminance regions according to their specific requirements. Dark regions receive fine quantization with higher effective precision to maintain luminance discrimination, while bright regions use coarser quantization, optimizing the balance between luminance precision and data amount for the entire image.
Data Source
AI summary
To decide an efficient tone-conversion characteristic by which it is possible to reduce tone loss while suppressing the occurrence of a visual tone gap, an image processing apparatus that decides a tone-conversion characteristic for converting an M level input image into an N level output image, the apparatus decides a maximum d at which a difference between a luminance value Y(i) corresponding to an ith level of the input image and a luminance value Y(i+d) corresponding to an (i+d)th level does not exceed a given threshold; and sets, based on the d decided by the decision unit, the tone-conversion characteristic so that the ith level and the (i+d)th level of the input image respectively correspond to a jth level and a (j+1)th level of the output image.