Neural Visual Coding With Integer Adjustment Factors

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Neural network-based image and video coding technologies face challenges in achieving optimal coding quality due to the use of floating-point numbers in adjustment factors, leading to susceptibility to coding errors and suboptimal performance.

Innovation Solution

The method involves determining adjustment factors using integer operations on integer parameters and reference factors to ensure device independence, thereby reducing coding errors and improving quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If floating-point numbers are used in adjustment factors for neural network-based coding, then coding flexibility is improved, but device independence deteriorates and coding errors increase

Engineering Contradiction:

Improvecoding flexibilityVSAvoiddevice independence

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent changes the data type parameter of adjustment factors from floating-point to integer. This parameter change resolves the contradiction by maintaining coding flexibility through integer arithmetic while ensuring device independence and eliminating floating-point-related coding errors across different processing devices.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If floating-point numbers are used in adjustment factors, then calculation precision is improved, but coding reliability deteriorates due to device-specific errors

Engineering Contradiction:

Improvecalculation precisionVSAvoidcoding reliability

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent substitutes the floating-point arithmetic mechanism with integer arithmetic mechanism. This replacement maintains calculation precision for the specific application context while eliminating device-specific errors caused by floating-point representation and operations, thereby improving coding reliability.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If integer operations are used for adjustment factors, then device independence is improved, but coding flexibility may be reduced

Engineering Contradiction:

Improvedevice independenceVSAvoidcoding flexibility

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent changes the data type parameter from floating-point to integer, which resolves the contradiction by demonstrating that integer arithmetic provides sufficient precision and flexibility for neural network-based coding while ensuring device independence and eliminating floating-point errors.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20250373827A1Method, apparatus, and medium for visual data processing

Publication Date: 2025.12.04 DOUYIN VISION CO LTD

AI summary

Embodiments of the present disclosure provide a solution for visual data processing. A method for visual data processing is proposed. The method comprises: determining, for a conversion between visual data and a bitstream of the visual data with a neural network (NN)-based model, an adjustment factor by performing one or more integer operations on at least one adjustment parameter and at least one reference factor, wherein the adjustment factor is used to adjust one or more samples associated with a latent representation of the visual data in an adjustment process of the NN-based model, and each of the at least one reference factor and the at least one adjustment parameter is an integer; and performing the conversion based on the adjustment factor.