In-Loop Temporal Pre-Filtering for Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video encoding architectures fail to adequately address temporal artifacts such as flicker or pulsation effects at GOP boundaries, which are prominent in low bit-rate video compression, especially in standards like MPEG-2 and H.264, due to the lack of consideration for encoding process artifacts.
Innovation Solution
An in-loop temporal pre-filtering method that combines current input video frames with previously reconstructed frames from the encoding process, using motion compensation and weighted averaging to reduce artifacts and enhance temporal correlation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If conventional spatial or temporal noise reduction schemes are used outside the encoder loop, then noise removal and temporal correlation enhancement are achieved, but encoding process artifacts such as temporal flicker and pulsation at GOP boundaries are not addressed
Solution Approach 1:
The patent introduces feedback by incorporating reconstructed pictures from the encoding process into the temporal filtering stage. The filter uses both the original input picture and the reconstructed picture as inputs, allowing the filtering process to adapt to encoding artifacts. This feedback loop enables the system to compensate for temporal flicker and pulsation artifacts that occur at GOP boundaries, while still maintaining noise removal capabilities.
Solution Approach 2:
The patent applies temporal filtering before the encoding process (pre-filtering), using reconstructed pictures from previous encoding iterations. By performing the filtering action in advance and using feedback from the encoding process, the system prepares the picture data to be more resistant to temporal artifacts while maintaining coding efficiency.
2Reliability
If conventional temporal filtering is applied to enhance temporal correlation, then noise removal is improved, but temporal artifacts at GOP boundaries persist due to lack of consideration for encoding artifacts
Solution Approach 1:
The patent uses feedback from the encoding process by incorporating reconstructed pictures into the temporal filtering. This allows the filter to adapt to the specific artifacts generated by the encoding process, improving temporal correlation while compensating for temporal flicker and pulsation artifacts at GOP boundaries.
Solution Approach 2:
The patent dynamically adjusts filtering parameters based on picture type (intra-coded vs. inter-coded) and encoding mode. By changing filtering parameters according to the specific encoding context, the system enhances temporal correlation where beneficial while avoiding the introduction or amplification of temporal artifacts in problematic regions.
3Reliability
If noise reduction is performed using conventional methods, then spatial and temporal correlation are improved, but important picture details such as edges and texture may be removed
Solution Approach 1:
The patent applies different filtering strengths based on local picture characteristics. By analyzing picture type and encoding mode, the filter adapts its behavior to preserve important local features such as edges and textures in regions where they are critical, while applying stronger filtering in regions where correlation enhancement is more beneficial.
Solution Approach 2:
The patent uses dynamic filtering that adapts to local picture characteristics and encoding context. The filtering strength and behavior change dynamically based on whether the picture is intra-coded or inter-coded, and based on local variance and detail content, thereby preserving important picture details while enhancing correlation where appropriate.
Data Source
AI summary
A method, system and device are provided for pre-filtering device for filtering a video signal prior to digitally encoding. The method includes receiving at least one input picture and at least one reconstructed picture from an encoding process and performing an in-loop temporal filtering process using at least one input picture and at least one reconstructed picture from an encoding process to output a pre-filtered video signal for use in an encoding process. The result is enabling an encoding process to produce an output with increased temporal correlation between adjacent pictures regardless of the coding type, since the artifacts introduced by the encoding process are also considered by the process.