Adaptive Prediction Filter for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing inter-prediction methods for video compression are inefficient in predicting target regions, leading to suboptimal coding efficiency due to the use of fixed displacement vectors and lack of effective noise reduction in motion compensated prediction.
Innovation Solution
A method is developed to design a prediction filter with novel parameterization for enhancing coding efficiency by selecting base filter kernels and determining modulation scalars to minimize a cost function, allowing for efficient prediction of target regions in video encoding and decoding processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If fixed displacement vectors are used in inter-prediction, then the prediction process is simple, but the coding efficiency is suboptimal
Solution Approach 1:
The patent applies dynamics by transitioning from fixed displacement vectors to adaptive prediction filters that dynamically adjust to local image characteristics. The prediction filter coefficients are modified based on the frequency characteristics of the target block, allowing the system to adapt to varying video content and achieve optimal coding efficiency while maintaining reasonable complexity through structured adaptation mechanisms.
2Reliability
If motion compensated prediction is used, then future frames can be predicted, but noise remains in the target image reducing coding efficiency
Solution Approach 1:
The patent applies parameter changes by modifying the prediction filter coefficients based on the frequency characteristics of the target block. Different filter coefficients are selected or adapted according to the frequency content analysis, which optimizes the prediction accuracy for different types of image regions. This parameter adaptation reduces noise in the predicted images while maintaining high coding efficiency through tailored prediction parameters for each block.
3Productivity
If prediction filters are designed with novel parameterization, then coding efficiency is enhanced, but the filter design complexity increases
Solution Approach 1:
The patent applies parameter changes by introducing a novel parameterization approach where filter coefficients are modified based on frequency characteristics. Instead of using complex custom filters for each block, the system adjusts a set of learnable parameters (filter coefficients) based on block frequency analysis. This parameter-based approach enhances coding efficiency while keeping the design manageable through systematic parameter modification rather than completely new filter synthesis.
Solution Approach 2:
The patent applies preliminary action by pre-calculating and storing a library of candidate filter coefficients or frequency characteristic profiles before actual prediction occurs. During encoding/decoding, the system quickly selects or interpolates from these pre-computed parameters based on the target block's frequency characteristics, avoiding the need for complex real-time filter design. This preliminary preparation significantly reduces operational complexity while maintaining high coding efficiency.
Data Source
AI summary
A method of encoding a video signal includes selecting a set of base filter kernels from a filter bank; determining a prediction filter parameter based on the set of base filter kernels; performing a filtering of a reference region for a target region based on the prediction filter parameter; and predicting the target region based on the filtered reference region, wherein the prediction filter parameter includes at least one of modulation scalar and partition information.