Video Filter Boundary Handling for Encoder-Decoder Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards lack a clear definition for handling pixels beyond picture or sub-picture boundaries during filtering operations, leading to inconsistent behavior between video encoders and decoders, which can result in poor video decoding quality.
Innovation Solution
Implementing techniques for picture and sub-picture boundary processing, where video encoders and decoders follow the same behavior by either not applying bilateral filtering (BIF) or Hadamard transform domain filtering (HTDF), applying padding to filter templates, or using partial filter templates and renormalizing outputs when pixels are required beyond boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If filtering operations (BIF or HTDF) are applied to blocks requiring pixels beyond picture or sub-picture boundaries, then filter accuracy is improved, but encoder-decoder consistency deteriorates due to lack of clear boundary handling definitions
Solution Approach 1:
The patent applies padding to filter templates before performing filtering operations near picture or sub-picture boundaries. This preliminary action provides the necessary pixel values beyond boundaries, enabling accurate filtering while ensuring both encoder and decoder handle boundary cases identically, thus maintaining consistency.
Solution Approach 2:
The patent introduces padding values as an intermediary element that mediates between the filter template and the picture boundary. These padding values allow the filtering operation to access pixels beyond boundaries without causing encoder-decoder inconsistency, effectively resolving the contradiction between filter accuracy and consistency.
2Measurement precision
If padding is applied to filter templates to enable filtering at boundaries, then filter accuracy is maintained, but device complexity increases
Solution Approach 1:
The patent changes the parameters of the filter template by applying padding values, transforming it from a fixed-size template to one that can adaptively extend beyond picture boundaries. This parameter change enables consistent boundary handling while maintaining filter accuracy through standardized padding procedures.
3Productivity
If transform and quantization are not applied to a block, then coding efficiency is improved, but filtering operations become unnecessary increasing processing overhead
Solution Approach 1:
The patent implements dynamic conditional filtering where the application of BIF or HTDF is determined based on whether transform and quantization were applied to the block. This dynamic approach allows the system to adaptively enable or disable filtering operations, optimizing processing overhead while maintaining coding efficiency.
Data Source
AI summary
An example device for coding video data includes a memory configured to store a block of video data and one or more processors, implemented in circuitry, and communicatively coupled to the memory. The one or more processors are configured to determine whether a transform and quantization are applied to the block of the video data and based on the transform and quantization not being applied to the block of video data, code the block of video data without applying bilateral filtering (BIF) or Hadamard transform domain filtering (HTDF).