DC Prediction Value Calculation Using Symmetric Reference Samples
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video codecs face challenges in accurately calculating the DC prediction for image blocks, particularly due to the complexity of averaging a large number of reference samples, which can result in poor prediction estimates and increased computational complexity.
Innovation Solution
A method is introduced to select a symmetric subset of reference samples from the top and left boundaries of an image block to determine a DC prediction value, which reduces computational complexity while improving prediction accuracy by centering the average coordinates closer to the block's center, thereby providing a better statistical estimate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large number of reference samples are used to calculate DC prediction, then prediction accuracy may improve, but computational complexity increases significantly
Solution Approach 1:
The patent extracts only the necessary subset of reference samples (specifically, samples from the top and left boundaries of the block) rather than using all available reference samples. This selective extraction reduces the number of samples involved in DC prediction calculation while maintaining adequate prediction accuracy, thereby resolving the contradiction between prediction accuracy and computational complexity.
Solution Approach 2:
The patent segments the reference samples into different groups based on their spatial location (top boundary samples and left boundary samples). By dividing the reference samples into these segments and selectively using only the boundary samples for DC prediction, the computational load is reduced while preserving the essential prediction information.
2Loss of information
If all reference samples are used for DC prediction, then more information is available for prediction, but the calculation becomes more complex and time-consuming
Solution Approach 1:
The patent extracts only the essential reference samples from the top and left boundaries of the block, discarding the redundant interior samples. This extraction process maintains the key prediction information needed for accurate DC prediction while significantly reducing the calculation time required.
3Measurement precision
If a symmetric subset of reference samples is selected, then prediction accuracy improves by centering coordinates, but sample selection complexity increases
Solution Approach 1:
The patent applies asymmetry by treating the top and left boundaries differently from the bottom and right boundaries. Specifically, it selects reference samples only from the top and left boundaries (where neighboring blocks exist) rather than all four boundaries. This asymmetric selection simplifies the sample selection process while achieving centered coordinates for improved prediction accuracy.
Data Source
AI summary
The disclosure relates to encoding and decoding image information. The encoding comprises receiving a block of pixels; determining a set of potential reference samples for the block of pixels; selecting a subset of the set of potential reference samples to be used as reference samples for the block of pixels; and using the selected reference samples to determine a DC prediction value for the block of pixels. A prediction error is determined for a pixel in the block of pixels on the basis of the DC prediction value. The decoding comprises receiving an encoded block of pixels; determining reference samples for the encoded block of pixels; and using the determined reference samples to define a DC prediction value for the block of pixels. A prediction error is received for a pixel of the encoded block of pixels. The pixel value is reconstructed on the basis of the DC prediction value.