Video Quality Assessment Using Intentional Distortion Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing no-reference video quality metrics cannot distinguish between intentional and unintentional distortions in content items, leading to erroneous quality scores.
Innovation Solution
Classify content items based on quality type, with original content items (first quality type) having intentional distortions and non-original content items (second quality type) having unintentional distortions, adjusting the weight of distortion parameters accordingly in video quality assessments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If no-reference video quality metrics are used to measure distortion levels, then quality assessment can be performed without reference content, but intentional distortions are incorrectly penalized leading to erroneous quality scores
Solution Approach 1:
The patent segments distortions into two categories: intentional distortions (e.g., film grain, vintage effects) and unintentional distortions (e.g., compression artifacts, noise). By classifying content as source material or processed material, the system applies different weighting strategies to different distortion types, allowing quality assessment to ignore intentional distortions while penalizing unintentional ones.
Solution Approach 2:
The patent applies local quality by using different quality assessment parameters and weights for different content types. Source material content receives one set of distortion weights that favor creative distortions, while processed material content receives another set that penalizes degradation artifacts more heavily. This localized approach to quality metrics resolves the contradiction between ease of operation and measurement precision.
2Device complexity
If all distortions are treated equally in quality assessment, then the assessment process is simple and uniform, but intentional distortions negatively impact quality scores causing misclassification
Solution Approach 1:
The patent implements dynamic distortion weighting where the assessment parameters change based on content classification. The system dynamically adjusts which distortion metrics are applied and their respective weights based on whether the content is identified as source material or processed material. This dynamic adaptation increases reliability without significantly increasing complexity, as the complexity is managed through automated content type detection.
Solution Approach 2:
The patent changes assessment parameters based on content type identification. Different sets of distortion weights are applied: source material uses weights that tolerate creative distortions, while processed material uses weights that emphasize degradation detection. This parameter changing strategy resolves the contradiction by making the assessment reliable across different content types while maintaining a unified assessment framework.
Data Source
AI summary
Methods and systems for determining content quality are disclosed. At least one distortion associated with a content item may be determined. Based on the content item comprising a source content item, it may be determined that the at least one distortion comprises at least one intentional artifact. A quality score associated with the content item may be determined. The at least one intentional artifact may have no effect on the quality score or a lower effect on the quality score than an unintentional artifact associated with the content item.