Intra Block Coding Using Pixel Group Prediction and Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards like MPEG-4 AVC limit prediction accuracy due to large spatial distances between predictor and predicted pixels, leading to inefficient coding and reliance on extrapolation instead of interpolation.
Innovation Solution
Divide pixels within a block into at least two groups, encode the first group before the second, using neighboring pixels for improved prediction, and apply interpolation to enhance coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pixels above or to the left of the current block are used as predictors in current coding standards, then the coding process is simple, but the prediction accuracy is limited due to large spatial distances
Solution Approach 1:
The current block is divided into multiple sub-blocks, and different prediction modes are applied to different sub-blocks. This segmentation allows the use of multiple reference pixel groups (above and left neighbors) for each sub-block, improving prediction accuracy for bottom-right pixels without excessively increasing overall coding complexity through selective application.
Solution Approach 2:
The patent extends the prediction approach by utilizing reference pixels from both horizontal (left) and vertical (above) dimensions simultaneously, rather than relying on a single direction. This multi-dimensional reference approach increases the number of available predictors and improves accuracy for pixels farther from the reference boundaries.
2Measurement precision
If extrapolation is used instead of interpolation due to causality limitations, then the coding process is straightforward, but the prediction accuracy deteriorates
Solution Approach 1:
The patent performs preliminary encoding of the current block's pixels in a specific sequence (e.g., row by row from top to bottom, or column by column from left to right) before final reconstruction. This preliminary action allows interpolated prediction values to be computed and stored as reference pixels for subsequent pixel predictions within the same block, enabling interpolation instead of extrapolation while maintaining coding simplicity.
Solution Approach 2:
The patent implements a feedback mechanism where predicted pixel values are fed back into the block as reconstructed reference pixels during the encoding process. This allows later pixels within the same block to use interpolated predictions from previously encoded pixels, improving accuracy while maintaining causality through the feedback loop.
3Measurement precision
If a single prediction mode is applied to the entire block, then the coding process is efficient, but the prediction accuracy for specific regions deteriorates
Solution Approach 1:
The current block is divided into multiple sub-blocks (e.g., 4×4 sub-blocks within a larger block), and different prediction modes are selectively applied to different sub-blocks based on local characteristics. This segmentation improves prediction accuracy for specific regions while maintaining coding efficiency through selective mode application and potential mode inheritance from parent blocks.
Solution Approach 2:
The patent applies different prediction strategies to different regions of the block based on local characteristics. For example, sub-blocks near the top use vertical reference pixels, while sub-blocks near the left use horizontal reference pixels, and bottom-right sub-blocks benefit from both. This local quality approach optimizes prediction accuracy for each region without requiring complete re-encoding of the entire block.
Data Source
AI summary
Methods and apparatus are provided for intra coding a block having pixels assigned to groups. An apparatus includes a video encoder for encoding a block in a picture using intra prediction by dividing pixels within the block into at least a first group and a second group and encoding the pixels in the first group prior to encoding the pixels in the second group. A prediction for at least one of the pixels within the second group is obtained by evaluating the pixels within the first group and the second group.