Intra Prediction Reference Padding for Wide Angular Video Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression technologies, such as HEVC, face challenges in efficiently handling wide angular intra prediction modes for non-square blocks, particularly in intra sample substitution and filtering processes, leading to inefficiencies in video encoding and decoding.
Innovation Solution
Implementing reference sample padding and filtering techniques for intra prediction, specifically adapting wide angular intra prediction modes for non-square blocks, and enhancing the intra prediction process to improve encoding and decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If wide angular intra prediction modes are applied for non-square blocks, then prediction accuracy is improved, but reference sample availability deteriorates
Solution Approach 1:
The patent applies preliminary action by performing reference sample padding before the intra prediction process. Extra reference samples are prepended or appended to the reference sample arrays in advance, ensuring that wide angular prediction modes have sufficient reference samples available without requiring additional complex processing during prediction.
Solution Approach 2:
The patent segments the reference sample processing into distinct stages: padding stage (adding extra samples), filtering stage (applying asymmetric filters), and prediction stage (using processed samples). This segmentation allows each stage to be optimized independently, with filtering applied only where needed based on prediction mode and block shape.
2Measurement precision
If reference sample padding is performed for wide angular prediction, then prediction performance is improved, but computational complexity increases
Solution Approach 1:
The patent applies local quality by using asymmetric filtering that adapts to the specific prediction mode and block shape. Different filtering operations are applied to different portions of the reference samples based on their relevance to the prediction direction. For example, more filtering is applied in directions where reference samples are more abundant, and less filtering where samples are scarcer.
Solution Approach 2:
The patent changes parameters dynamically based on prediction mode and block dimensions. The filtering strength, filter length, and padding amount are adjusted according to the specific wide angular mode being used and whether the block is wider or taller. This parameter adaptation optimizes performance while controlling complexity for each specific case.
3Measurement precision
If asymmetric filtering is applied to reference samples, then intra prediction accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies partial action by selectively filtering only the necessary reference samples rather than all reference samples uniformly. The filtering is applied partially based on the prediction mode and block shape, processing only the samples that will actually be used in the prediction calculation. This reduces unnecessary processing time while maintaining accuracy.
Solution Approach 2:
The filtering operation is performed as a preliminary step before the actual intra prediction calculation. By pre-filtering the reference samples, the computationally intensive filtering is done once, and the results are reused during prediction, rather than performing filtering repeatedly during the prediction process itself.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of and an apparatus for controlling intra prediction for decoding of a video sequence are provided. The method includes setting a table including a plurality of intra prediction modes respectively corresponding to a plurality of intra prediction angles, the plurality of intra prediction angles including diagonal intra prediction angles respectively corresponding to diagonal directions of all shapes of coding units, and selecting, for decoding the video sequence, one of the plurality of intra prediction modes, from the set table.