Neural Network Post-Filter Energy Signaling in Video Decoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Modern video coding systems face challenges in managing energy consumption, particularly with the introduction of neural-network post-filters that increase energy requirements due to their complexity and the processing of high-resolution and high-dynamic range video content, which is not aligned with the global need for reduced energy consumption in the entertainment industry.

Innovation Solution

Incorporating syntax elements in the video bitstream to indicate energy parameters related to neural network post-processing filters, allowing decoders to make informed decisions about applying these filters based on energy criteria, thereby optimizing energy usage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If neural network post-processing filters are applied to improve video quality, then video processing quality is improved, but energy consumption increases

Engineering Contradiction:
Improvevideo processing qualityVSAvoidenergy consumption
Core Design Contradiction:
Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The patent implements dynamic selection of neural network post-processing filters based on energy criteria. The system adapts its processing approach by selecting different filter configurations or intensities according to available energy resources, allowing quality optimization when energy is abundant and energy conservation when resources are limited.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent introduces energy parameter syntax elements that encode information about the energy consumption characteristics of different neural network filter configurations. By transmitting and utilizing these parameters, the system can make informed decisions about filter application, adjusting processing intensity and selection based on energy constraints while maintaining quality where possible.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If complex neural network post-filters are used to enhance video processing, then processing capability is improved, but device complexity increases

Engineering Contradiction:
Improveprocessing capabilityVSAvoiddevice complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the neural network post-processing functionality into multiple filter options with different complexity levels and energy consumption characteristics. Instead of implementing a single complex filter, the system provides a library of filters that can be selectively applied, allowing the decoder to choose the appropriate level of processing complexity based on device capabilities and energy constraints.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system dynamically selects from multiple filter configurations based on runtime conditions including energy availability and device capabilities. This dynamic selection mechanism allows the system to optimize between processing capability and device complexity by choosing appropriate filter configurations rather than always applying the most complex processing.

Inventive Principle:
Principle #15Dynamics

3Manufacturing precision

If high-resolution and high-dynamic range video content is processed, then video quality is improved, but energy consumption increases

Engineering Contradiction:
Improvevideo qualityVSAvoidenergy consumption
Core Design Contradiction:
Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The patent encodes energy parameter syntax elements in the bitstream that provide information about the energy consumption characteristics of processing different video configurations. This allows the decoder to assess whether the available energy resources support processing of high-resolution and high-dynamic range content, and to adjust processing parameters accordingly to balance quality and energy consumption.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The system dynamically adjusts its processing of high-resolution and high-dynamic range video content based on energy criteria. When energy resources are sufficient, the system applies full-quality processing; when energy is constrained, the system can reduce processing intensity or skip certain processing steps, thereby adapting to energy conditions while maintaining quality where possible.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4676043A1Energy information for neural network post filter
Publication Date: 2026.01.07 INTERDIGITAL CE PATENT HOLDINGS SAS
  • EP4676043A1 patent drawingFigure 1~2
  • EP4676043A1 patent drawingFigure 3
  • EP4676043A1 patent drawingFigure 4

AI summary

Specific syntax elements define energy characteristics of a neural network post-processing filter (NNPF) used in a video coding system. An encoded video bitstream carries such syntax elements from an encoding device to a decoding device, thus allowing the encoding device to provide energy consumption information related to the NNPF and allowing the decoding device to handle its energy consumption by deciding whether or not it should apply the NNPF. An encoding method, a decoding method, an encoding apparatus and a decoding apparatus are described.