Enhanced MOS Predictive Model for HAS Video QoE Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for evaluating the quality of experience (QoE) of HTTP adaptive streaming (HAS) video services are inadequate as they do not account for bit rate fluctuations, which significantly impact user satisfaction, unlike traditional video services where bit rates are constant.
Innovation Solution
The method involves pre-processing the peak signal to noise ratio (PSNR) of each video segment at different bit rates to calculate differential PSNR (dPSNR) values, and using these along with mean, maximum, and standard deviation values to create an enhanced PSNR predictive model, which is then used to determine an enhanced mean opinion score (MOS) predictive model, considering various factors affecting video quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional MOS evaluation methods using mean PSNR or mean SSIM are applied to HAS video services, then the evaluation process remains simple, but the measurement accuracy of QoE deteriorates because bit rate fluctuations are not considered
Solution Approach 1:
The patent transforms the single mean PSNR parameter into multiple parameters including mean PSNR, standard deviation of PSNR, and bit rate fluctuation metrics. This parameter expansion allows the evaluation model to capture both quality and stability aspects of HAS video services, resolving the contradiction between measurement accuracy and model simplicity.
Solution Approach 2:
The patent adds a new dimension to the evaluation by incorporating bit rate fluctuation characteristics alongside traditional quality metrics. This dimensional expansion enables the model to assess QoE comprehensively by considering both video quality and network adaptation performance, thereby improving measurement accuracy without excessive complexity.
2Measurement precision
If subjective MOS scoring tests are conducted in strict test environments following standard procedures, then measurement accuracy improves, but the ease of operation deteriorates due to high environmental requirements and complex procedures
Solution Approach 1:
The patent creates an objective evaluation model that copies and simulates the outcomes of subjective MOS scoring through mathematical relationships between objective parameters (PSNR, SSIM, bit rate fluctuations) and subjective quality perceptions. This copying approach eliminates the need for actual human scoring while maintaining measurement accuracy, thus resolving the contradiction between precision and operational ease.
Solution Approach 2:
The patent replaces the mechanical system of human subjective scoring with an automated objective evaluation system based on mathematical models and algorithmic computations. This substitution eliminates environmental constraints and procedural complexity associated with human testing while preserving the ability to measure QoE accurately.
3Ease of operation
If only mean PSNR values are considered for QoE evaluation, then the ease of operation improves, but measurement precision deteriorates because bit rate fluctuation status is ignored
Solution Approach 1:
The patent performs preliminary calculations of multiple parameters (mean PSNR, standard deviation of PSNR, bit rate fluctuation metrics) before the actual QoE evaluation. This preliminary action prepares comprehensive data that captures both quality and stability characteristics, enabling accurate QoE measurement while maintaining operational simplicity during the evaluation phase.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
Embodiments of the present invention provide a method and an apparatus for measuring quality of experience of a mobile video service. The method includes: processing a PSNR of each segment of each sample video, determining an ePSNR predictive model according to preset parameters obtained after processing and mean opinion scores of all sample videos, and determining an enhanced mean opinion score MOS predictive model according to the predictive model. Then, for any video that needs to be evaluated, QoE of the video that needs to be evaluated may be determined according to only the enhanced MOS predictive model and an ePSNR determined according to the ePSNR predictive model. In comparison with a prior-art method for determining QoE in which only a mean value of PSRNs of all frames is considered, in this process of measuring quality of experience of a mobile video service, as many as factors that affect a PSNR of a video are considered. Therefore, accurate measurement of quality of experience of an HAS video service can be implemented.