Intra Prediction Smoothing Filter for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in improving compression efficiency and reducing computational load, particularly in intra prediction methods where outliers in reference image samples can negatively affect prediction results and increase computational overhead.
Innovation Solution
The use of a smoothing filter to one-dimensionally filter neighboring image samples along specific rays, determining prediction values from filtered sequences, and performing intra prediction with increased precision to reduce residual signals and computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional intra prediction methods are used without smoothing filters, then the prediction process is simpler and faster, but outliers in reference image samples negatively affect prediction results and reduce compression efficiency
Solution Approach 1:
The smoothing filter is applied to reference image samples before they are used in the intra prediction process. This preliminary filtering action removes outliers and smooths the reference data, ensuring that the subsequent prediction operations work with cleaned input data, thereby improving prediction accuracy without adding complex real-time processing during prediction
Solution Approach 2:
The smoothing filter acts as an intermediary component between the reference image samples and the intra prediction algorithm. It mediates the quality of reference data by filtering out outliers before they can negatively impact the prediction process, serving as a buffer that improves overall system reliability
2Measurement precision
If higher precision is used in intra prediction calculations, then prediction accuracy improves, but computational load increases
Solution Approach 1:
The patent applies precision scaling by representing filter coefficients and intermediate calculation results in fixed-point arithmetic with specific precision parameters. By carefully selecting precision levels for different stages of calculation and using scalable precision modes, the system achieves high prediction accuracy when needed while allowing lower precision operations for simpler cases, thereby managing computational load effectively
3Manufacturing precision
If smoothing filters are applied to all reference samples, then prediction quality improves, but processing time and computational overhead increase
Solution Approach 1:
The patent implements selective smoothing where the filter is applied only to reference samples that are likely to contain outliers or that significantly impact prediction accuracy. By identifying and filtering only the necessary portion of reference data rather than processing all samples uniformly, the system achieves improved prediction quality with reduced processing time and computational overhead
Data Source
AI summary
An apparatus for intra predicting a block to be predicted of a picture from neighboring image samples sequentially extending along two rays starting at a common intersection point and pointing into different directions, the neighboring image samples including edge neighboring samples bordering the block to be predicted, includes a smoothing filter and a determiner. The smoothing filter is configured to one-dimensionally filter the neighboring image samples, which are sequentially arranged into a sequence by anti-parallelly traveling along a first one of the two rays and parallelly traveling along a second one of the two rays to obtain a filtered image sample sequence. The determiner is configured to determine a prediction value for a predetermined image sample of the block to be predicted by using the filtered image sample sequence.