Tunable VMAF Scoring With Enhancement Gain Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The impact of image enhancement operations on perceptual video quality estimations is difficult to assess accurately using VMAF scores when these operations are applied within a codec, leading to erroneous attribution of enhancement gains to data compression operations, which complicates codec evaluation.
Innovation Solution
A tunable VMAF application is used to compute tuned VMAF scores by limiting enhancement gains through configurable VIF and DLM limits, reducing the influence of image enhancement operations on perceptual video quality estimates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If image enhancement operations are applied within a codec to improve visual quality, then perceptual video quality is improved, but the impact of data compression operations on quality becomes difficult to assess accurately
Solution Approach 1:
The patent segments the quality assessment process into two distinct components: enhancement gain measurement and compression quality measurement. By separating these previously conflated measurements, the system can independently evaluate the contribution of image enhancement operations versus data compression operations, thereby resolving the contradiction between improving overall quality and maintaining accurate measurement of compression impact.
2Reliability
If VMAF scores are used to evaluate codecs with image enhancement operations, then overall quality is captured, but enhancement gains are erroneously attributed to data compression operations
Solution Approach 1:
The patent extracts the enhancement gain component from the overall VMAF score by computing it separately using reference videos processed with and without enhancement operations. This extraction allows the system to isolate and remove the enhancement contribution from the quality measurement, enabling accurate attribution of quality metrics specifically to data compression operations rather than conflating them with enhancement effects.
3Illumination intensity
If enhancement gains are not limited in VMAF scoring, then visual quality appears improved, but codec evaluation becomes inaccurate
Solution Approach 1:
The patent changes the parameter of enhancement gain limitation by introducing configurable limits on the amount of enhancement gain that can be attributed to quality improvements. By adjusting this parameter, the system can control the influence of enhancement operations on VMAF scores, allowing accurate codec evaluation while still accounting for reasonable enhancement contributions to visual quality.
Data Source
AI summary
In various embodiments, a tunable VMAF application reduces an amount of influence that image enhancement operations have on perceptual video quality estimates. In operation, the tunable VMAF application computes a first value for a first visual quality metric based on reconstructed video content and a first enhancement gain limit. The tunable VMAF application computes a second value for a second visual quality metric based on the reconstructed video content and a second enhancement gain limit. Subsequently, the tunable VMAF application generates a feature value vector based on the first value for the first visual quality metric and the second value for the second visual quality metric. The tunable VMAF application executes a VMAF model based on the feature value vector to generate a tuned VMAF score that accounts, at least in part, for at least one image enhancement operation used to generate the reconstructed video content.