Encrypted Video Quality Scoring via Network Statistics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for evaluating video quality in impaired network video quality monitoring (IN VQM) are limited as they require direct access to video content and are costly and time-consuming, especially when dealing with encrypted video streams and diverse device and service pairings, and do not accurately predict user experience across different network conditions.
Innovation Solution
A system that uses non-reference perceptual video quality analysis (NR VQA) scores and network statistics to train an impaired network video quality analyzer (IN VQA) without decrypting the video stream, allowing for the evaluation of video delivery quality on live networks using synthetic impairment generation and machine learning techniques, enabling the creation of training sets that correlate IP network behavior with device and video source pairs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If subjective human evaluation panels are used to determine video quality scores, then accurate perceptual video quality assessment is achieved, but the process becomes cumbersome and expensive
Solution Approach 1:
The patent replaces the mechanical system of human evaluation panels with an automated computer-based scoring system that uses machine learning models trained on network statistics and video delivery parameters. This substitution eliminates the need for human observers while maintaining measurement accuracy through algorithms that correlate network conditions with perceived video quality.
Solution Approach 2:
The system enables self-service by allowing automated collection and analysis of video quality metrics without requiring human intervention. The machine learning models automatically process network statistics, video delivery data, and performance metrics to generate quality scores, making the evaluation process autonomous and eliminating the need for cumbersome human panels.
2Measurement precision
If video streams are decrypted for quality analysis, then direct access to video content is achieved, but the process becomes time-consuming and resource-intensive
Solution Approach 1:
The patent extracts the essential quality assessment function from the video content itself by analyzing network statistics and delivery parameters instead of requiring direct access to decrypted video frames. This extraction allows quality evaluation to be performed on encrypted streams by measuring network-level metrics that correlate with perceived quality, eliminating the need for time-consuming decryption and frame-by-frame analysis.
Solution Approach 2:
The system introduces network statistics and video delivery parameters as intermediary measurements that bridge the gap between encrypted video streams and quality assessment. These intermediaries provide indirect but accurate quality information without requiring decryption, enabling fast analysis by measuring network conditions that directly impact video delivery quality.
3Ease of operation
If current network video quality models are used, then network component operation is analyzed, but user subjective experience of video quality cannot be evaluated
Solution Approach 1:
The patent changes the parameters being measured from simple network component metrics to a comprehensive set of network statistics and video delivery parameters that collectively predict user experience. By transforming individual network metrics into a multi-dimensional parameter space that includes video quality scores, the system bridges the gap between objective network analysis and subjective user perception.
Solution Approach 2:
The system implements feedback loops where machine learning models are trained on correlated network statistics and video quality outcomes, continuously improving their ability to predict user experience. The feedback mechanism allows the system to learn from actual user perceptions and adjust its predictions, transforming static network analysis into a dynamic user experience evaluation system.
4Productivity
If automated collection of video quality metrics is implemented, then cost and time constraints are reduced, but the system requires sophisticated machine learning training and data correlation
Solution Approach 1:
The patent applies preliminary action by pre-training machine learning models with extensive network statistics and video delivery data before deployment. This pre-training phase creates a robust foundation that enables the system to automatically evaluate video quality without requiring complex real-time processing, transferring the computational complexity to an offline training stage rather than online operation.
Data Source
AI summary
At least three uses of the technology disclosed are immediately recognized. First, a video stream classifier can be trained that has multiple uses. Second, a trained video stream classifier can be applied to monitor a live network. It can be extended by the network provider to customer relations management or to controlling video bandwidth. Third, a trained video stream classifier can be used to infer bit rate switching of codecs used by video sources and content providers. Bit rate switching and resulting video quality scores can be used to balance network loads and to balance quality of experience for users, across video sources. Balancing based on bit rate switching and resulting video quality scores also can be used when resolving network contention.