Video Quality Indicators From Gradients and Lateral Averages

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for calculating video quality information are inadequate in terms of accuracy and efficiency, particularly in assessing the subjective quality of video signals transmitted via telecommunication networks.

Innovation Solution

A device and method that utilize non-linear transformations based on absolute gradient values and lateral average values of video frames to calculate video quality indicators, incorporating formulas like R = (max(0, (E-S))/(E+S+c(E)) and c(E) = 0.5 * (c0 + mean(E)) to enhance accuracy and efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If existing methods for calculating video quality information are used, then the calculation process is simple, but the accuracy and efficiency of video quality assessment is insufficient

Engineering Contradiction:
Improvevideo quality assessment accuracyVSAvoidcalculation process complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the video quality assessment process into distinct computational stages: gradient calculation, lateral average computation, non-linear transformation application, and final quality indicator derivation. This segmentation allows each component to be optimized independently while maintaining overall accuracy and efficiency.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transforms the video quality assessment by changing parameters from direct pixel comparisons to gradient-based features, then to lateral average values, and finally applying non-linear transformations. This parameter transformation sequence significantly improves measurement precision by capturing perceptually relevant features.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If existing methods for calculating video quality information are used, then the computational resources required are low, but the efficiency of video quality assessment is insufficient

Engineering Contradiction:
Improvevideo quality assessment efficiencyVSAvoidcomputational resource consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent extracts only the most relevant features for quality assessment: absolute gradient values and lateral average values. By taking out and focusing on these specific features rather than processing all video data, the method achieves high efficiency with reduced computational resource consumption.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs preliminary computations of gradient values and lateral averages before applying the final non-linear transformation. This preliminary action organizes and pre-processes the data in an efficient manner, reducing the computational burden during the actual quality assessment phase.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If simple calculation methods are used, then the computational load is low, but the alignment with human perception is insufficient

Engineering Contradiction:
Improvealignment with human perceptionVSAvoidcalculation method complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies non-linear transformations to the gradient and lateral average values to better align with human perceptual characteristics. This parameter change from linear to non-linear domain captures the complex ways humans perceive video quality, including sensitivity to different types of distortions.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent computes quality indicators locally at each pixel position using gradient and lateral average values, then aggregates these local measurements. This local quality approach allows the assessment to capture spatially varying quality characteristics that correlate with human perception of different regions in the video.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3757871B1Device and corresponding method for calculating video quality information of a video sequence
Publication Date: 2025.10.15 ROHDE & SCHWARZ GMBH & CO KG
  • EP3757871B1 patent drawingFigure 1
  • EP3757871B1 patent drawingFigure 2
  • EP3757871B1 patent drawingFigure 3

AI summary

A device (10) for calculating video quality information of a video sequence is provided. The device (10) comprises a data acquisition unit (11) for acquiring at least one video frame with respect to the video sequence, and a calculation unit (12) connected to the data acquisition unit (11). In this context, the calculation unit (12) is configured to calculate at least one video quality indicator value on the basis of a non-linear transformation with the aid of at least one absolute gradient value with respect to the at least one video frame and at least one lateral average value with respect to the at least one video frame.