Intra Prediction Reference Filtering for Lower-Residual Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression techniques face challenges in maintaining high video quality while minimizing data transmission rates and reducing residual block high-frequency components due to diverse reference blocks and pixel value differences at block boundaries, leading to decreased coding efficiency.
Innovation Solution
A method for decoding moving pictures that involves reconstructing intra-prediction blocks by generating unavailable reference pixels from available ones, adaptively filtering these pixels based on block size and mode, and using them to create prediction blocks, thereby minimizing residual signal data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If intra-prediction is performed using reference pixels from multiple adjacent reference blocks, then the prediction block can be generated for diverse block sizes, but pixel value differences at block boundaries generate high-frequency components in residual blocks, decreasing coding efficiency
Solution Approach 1:
The patent applies local quality by differentiating the treatment of reference pixels based on their position. Reference pixels at block boundaries (edge regions) are filtered differently from interior pixels. Specifically, a filtering process is applied to reference pixels at boundaries to smooth pixel value differences, while interior pixels are used without filtering. This localized differentiation resolves the contradiction by maintaining adaptability for diverse block sizes while reducing high-frequency components at boundaries that harm coding efficiency.
Solution Approach 2:
The patent changes the parameter of reference pixel values by applying a filtering operation to selected reference pixels. The filtering modifies the pixel values at block boundaries to reduce abrupt transitions. This parameter change allows the system to maintain versatility in handling diverse block sizes while improving coding efficiency by reducing the high-frequency components that arise from unfiltered boundary pixel differences.
2Ease of operation
If reference pixels are used directly without filtering, then the decoding process is simple and fast, but pixel value differences at block boundaries create high-frequency components in residual blocks, reducing coding efficiency
Solution Approach 1:
The patent applies local quality by selectively filtering only the reference pixels at block boundaries while leaving interior pixels unfiltered. This localized filtering approach maintains decoding simplicity for the majority of pixels while addressing the high-frequency component issue only where it occurs (at boundaries). The filtering is applied conditionally based on pixel position, preserving ease of operation while improving coding efficiency.
Solution Approach 2:
The patent applies partial action by filtering only a subset of reference pixels (those at block boundaries) rather than all reference pixels. This selective filtering maintains the simplicity of the decoding process for most pixels while applying the necessary filtering only where boundary effects occur. The partial application of filtering resolves the contradiction by balancing decoding simplicity with coding efficiency improvement.
3Adaptability or versatility
If multiple adjacent reference blocks are used for prediction, then diverse block sizes can be supported, but the difference of pixel values between reference pixels at both boundaries increases, leading to more high-frequency components in residual blocks
Solution Approach 1:
The patent applies local quality by differentiating the treatment of reference pixels based on their spatial location. Reference pixels at block boundaries undergo filtering to reduce pixel value differences, while interior pixels are used directly. This localized differentiation maintains support for diverse block sizes while improving prediction accuracy by reducing the boundary effects that cause high-frequency components and prediction errors.
Solution Approach 2:
The patent applies preliminary anti-action by pre-filtering the reference pixels at block boundaries before they are used in the intra-prediction process. This preliminary filtering action counteracts the potential harm of pixel value differences at boundaries before they can affect the prediction accuracy. By applying this anti-action in advance, the patent maintains versatility for diverse block sizes while preventing the degradation of prediction accuracy that would otherwise occur.
Data Source
AI summary
A method for encoding an image, the method including determining an intra prediction mode of a current block, generating a prediction block of the current block according to the intra prediction mode, generating a residual block by calculating a difference between the current block and the prediction block, transforming the residual block to generate a transform block, quantizing the transform block to generate a quantized block, and entropy-coding coefficients of the quantized block, wherein the generating the prediction block of the current block according to the intra prediction mode includes generating reference pixels on an unavailable position if there exist unavailable reference pixels of the current block based on the intra prediction mode, determining whether reference pixels of the current block are filtered or not based on the intra prediction mode and a block size of the current block, and filtering the reference pixels of the current block using a filter if it is determined that the reference pixels of the current block are filtered, wherein the reference pixels are adaptively filtered based on the block size of the current block, wherein the filter is not applied if it is determined that the size of the current block is smaller than a first size, wherein as the size of the current block is increased within the first size and a second size, a number of intra prediction modes to which the filter is applied is increased, wherein the filter is not applied to the intra prediction mode corresponding to a DC mode, and wherein the intra prediction mode is selected among a predetermined number of directional intra prediction modes and two non-directional intra prediction modes.