Video Quality Assessment Using Critical Quantization Parameter
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for assessing video quality are inaccurate due to the non-linear relationship between subjective quality and quantization parameter, as the subjective quality of a video does not change strictly monotonically relative to the quantization parameter used during encoding.
Innovation Solution
A method and apparatus that determine an actual and critical quantization parameter for a target video, where the critical quantization parameter is the maximum parameter at which distortion is not identifiable by human eyes, allowing for accurate assessment of video quality by comparing the actual parameter to the critical parameter.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If subjective quality is calculated according to quantization parameter using linear relationship assumption, then calculation is simple, but video quality assessment accuracy is poor
Solution Approach 1:
The patent introduces a critical quantization parameter as an additional parameter to transform the simple linear model into a piecewise linear model. By changing the parameter representation from single quantization parameter to pair of (actual quantization parameter, critical quantization parameter), the model captures the non-monotonic relationship between subjective quality and quantization parameter while maintaining calculation simplicity.
Solution Approach 2:
The patent segments the quantization parameter range into multiple regions based on the critical quantization parameter. When actual quantization parameter is less than critical quantization parameter, one quality assessment rule applies; when equal or greater, another rule applies. This segmentation allows the model to handle different quality degradation patterns in different quantization ranges, improving accuracy without complex computations.
2Device complexity
If linear relationship between subjective quality and quantization parameter is assumed, then assessment method is simple, but it cannot accurately reflect non-monotonic quality changes
Solution Approach 1:
The patent makes the quality assessment dynamic by introducing the critical quantization parameter that adapts to different video content and conditions. The assessment method dynamically switches between different calculation modes based on the relationship between actual and critical quantization parameters, allowing it to adapt to non-monotonic quality changes while keeping the underlying calculation rules simple and deterministic.
Data Source
AI summary
Embodiments of the present invention provide a method and an apparatus for assessing video quality. The method includes: determining an actual quantization parameter and a critical quantization parameter of the target video according to a target video, where the critical quantization parameter is a maximum quantization parameter of the target video in a case in which a distortion is not evidently identifiable by human eyes; and determining quality of the target video according to the actual quantization parameter and the critical quantization parameter. According to the method and apparatus of the present invention, a critical quantization parameter of a target video is determined, and the actual quantization parameter is compared with the critical quantization parameter, which can accurately determine whether subjective quality of the target video changes strictly monotonically relative to the used actual quantization parameter, so that the video quality can be assessed accurately.