Video Codec Evaluation with Constrained Cubic Interpolation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video codec evaluation tools are limited by the requirement for monotonically increasing input data and overlapping data sets, leading to unstable and unreliable performance evaluations, particularly when data patterns do not conform to these conditions.

Innovation Solution

A method involving curve fitting with constrained cubic interpolation is employed to analyze monotonicity conditions, allowing for the fitting of both monotonic and non-monotonic data sets, enabling accurate evaluation of codec performance by constructing fitting algorithms that respect the inherent data patterns of anchor and test codecs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If traditional evaluation tools require monotonically increasing input data and overlapping data sets, then the evaluation process is simplified, but the reliability and accuracy of codec performance evaluation deteriorates when data patterns do not conform to these conditions

Engineering Contradiction:

Improveevaluation process simplicityVSAvoidcodec performance evaluation reliability

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The patent transforms the evaluation approach by changing the mathematical parameters from requiring monotonicity to using curve fitting with cubic interpolation. This allows the system to handle non-monotonic data patterns while maintaining evaluation reliability through polynomial fitting that can accommodate various data behaviors without requiring strict monotonic increasing conditions.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

Instead of requiring data to conform to strict monotonic patterns, the patent inverts the approach by using curve fitting to adapt to whatever data patterns emerge. The cubic interpolation method can handle both monotonic and non-monotonic data, effectively inverting the traditional requirement structure.

Inventive Principle:

Principle #13The other way round (Inversion)

2Measurement precision

If curve fitting with constrained cubic interpolation is employed to analyze monotonicity conditions, then the accuracy of codec performance evaluation improves, but the device complexity and computational requirements increase

Engineering Contradiction:

Improvecodec performance evaluation accuracyVSAvoidevaluation system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces curve fitting and cubic interpolation as intermediary mathematical tools between the raw codec performance data and the final evaluation results. These intermediaries smooth out data irregularities and provide a continuous representation that improves measurement precision while managing the complexity through well-established mathematical methods.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces traditional mechanical or algorithmic data processing methods with mathematical curve fitting approaches. Instead of using complex procedural algorithms to handle monotonicity checks and data validation, the system uses cubic interpolation mathematics to automatically handle these requirements, reducing procedural complexity while improving precision.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Adaptability or versatility

If the evaluation method accommodates both monotonic and non-monotonic data sets, then the adaptability and versatility of the evaluation tool improves, but the difficulty of detecting and measuring codec performance characteristics increases

Engineering Contradiction:

Improveevaluation tool versatilityVSAvoidcodec performance measurement difficulty

Core Design Contradiction:

Adaptability or versatilityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent creates a universal evaluation method using cubic interpolation that can handle multiple data types (monotonic, non-monotonic, overlapping, non-overlapping) through a single unified approach. This multi-functional capability allows the same curve fitting algorithm to adapt to various data patterns without requiring separate processing paths, thereby improving versatility while managing measurement complexity.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS12452403B2Methods and devices for video codec performance measurement and evaluation

Publication Date: 2025.10.21 TENCENT AMERICA LLC

AI summary

This disclosure relates generally to video coding/decoding and particularly for evaluating codec performance. One method includes obtaining a first plurality of anchor data points; obtaining a second plurality of test data points; analyzing monotonicity conditions of the first plurality of anchor data points and the second plurality of test data points; determining a fitting algorithm based on the analyzed monotonicity conditions; fitting the first plurality of anchor data points with the fitting algorithm to obtain an anchor curve; fitting the second plurality of test data points with the fitting algorithm to obtain a test curve; and evaluating the test codec performance based on the anchor curve and the test curve, to obtain an evaluation result.