Neural Network Video Filtering with Bit-Depth Signaling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies, such as MPEG-2, MPEG-4, ITU-T.263, ITU-T.264/MPEG-4 AVC, ITU-T.265 HEVC, and VVC, require improvements in coding efficiency and effectiveness.

Innovation Solution

Implementing a neural network filter for video processing that includes indications of bit depths, patch sizes, and usage of color components in the bitstream to enhance coding efficiency and effectiveness.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional video coding technologies (MPEG-2, MPEG-4, HEVC, VVC) are used, then video compression is achieved, but coding efficiency and effectiveness need further improvement

Engineering Contradiction:
Improvecoding efficiencyVSAvoidcoding effectiveness
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent replaces traditional mechanical/video coding systems with a neural network-based filter system. The neural network filter processes video data through learned representations, substituting conventional coding mechanisms with AI-driven processing to achieve superior coding efficiency and effectiveness.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces bit depth parameters and patch size parameters as controllable variables in the neural network filter processing. By adjusting these parameters (e.g., different bit depths for color components, varying patch sizes), the system optimizes coding performance for different video characteristics and requirements.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If neural network filter is applied with specific bit depths and patch sizes, then coding efficiency improves, but system complexity increases

Engineering Contradiction:
Improvecoding efficiencyVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent implements dynamic parameter adjustment where the neural network filter can adaptively select bit depths and patch sizes based on video content characteristics. This dynamic behavior allows the system to optimize coding efficiency for each specific video segment while managing computational complexity through content-aware adaptation.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system performs preliminary analysis of video data to determine optimal bit depth and patch size parameters before applying the neural network filter. This preliminary action enables efficient parameter selection that balances coding performance with computational requirements, avoiding unnecessary complexity for simple video content.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS20250343907A1Method, apparatus, and medium for video processing
Publication Date: 2025.11.06 DOUYIN VISION CO LTD
  • US20250343907A1 patent drawing
  • US20250343907A1 patent drawing
  • US20250343907A1 patent drawing

AI summary

Embodiments of the present disclosure provide a solution for video processing. A method for video processing is proposed. In the method, a conversion between a current video unit of a video and a bitstream of the video is performed. A neural network filter is applied to the current video unit. An output of the neural network filter includes a set of color components. At least one indication indicating a set of bit depths of the set of color components is included in the bitstream.