Video Encoder Reference Sample Projection Smoothing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video encoding technologies, such as H.265, face challenges in improving the accuracy of reference samples for intra prediction, leading to suboptimal coding efficiency due to the introduction of undesirable frequencies and complexity in interpolation processing.
Innovation Solution
The proposed solution involves an encoder and decoder that derive and smooth one-dimensional arrays of reference samples by projecting decoded pixels onto a perpendicular line, applying spatial smoothing filters, and using interpolation filters regardless of integer positions, to enhance prediction image accuracy and simplify processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional interpolation processing is used to derive reference samples, then the prediction image can be generated, but undesirable frequencies are introduced and manufacturing precision deteriorates
Solution Approach 1:
The patent extracts only the necessary reference sample values from the decoded pixel data by projecting onto a second line perpendicular to the first line, rather than using all pixel values in conventional interpolation. This extraction approach derives the minimal set of reference samples needed for accurate prediction while avoiding the introduction of undesirable frequencies through selective sampling.
Solution Approach 2:
The patent performs preliminary smoothing on the one-dimensional array of reference samples before using them for prediction. This preliminary action removes undesirable frequencies and smooths the reference samples in advance, ensuring that the prediction image is generated from clean, frequency-optimized data without introducing harmful artifacts during the prediction process.
2Productivity
If complex interpolation processing is applied to improve prediction accuracy, then coding efficiency improves, but device complexity increases
Solution Approach 1:
The patent segments the reference sample derivation process into distinct steps: projecting decoded pixel values onto a second line to derive reference samples, smoothing the one-dimensional array, and then using these smoothed samples for prediction. This segmentation simplifies the overall process by breaking down complex interpolation into manageable operations, improving coding efficiency without excessive complexity.
Solution Approach 2:
The patent changes the dimensional approach by projecting pixel values from a first line onto a second line perpendicular to it, creating a one-dimensional array of reference samples. This dimensional transformation simplifies the interpolation process compared to conventional two-dimensional methods, reducing computational complexity while maintaining or improving prediction accuracy and coding efficiency.
3Manufacturing precision
If conventional reference sample derivation is used, then the process is simple, but prediction image accuracy deteriorates
Solution Approach 1:
The patent applies preliminary smoothing to the one-dimensional array of reference samples before generating the prediction image. This preliminary action enhances the accuracy of the prediction image by ensuring that the reference samples are smooth and free from undesirable frequencies, while the overall process remains relatively simple and does not require excessive computational complexity.
Data Source
AI summary
An encoder includes circuitry and memory. The circuitry, using the memory: derives a one-dimensional array of a plurality of reference samples for intra prediction; performs smoothing on the one-dimensional array of the plurality of reference samples which has been derived; and generates a prediction image using the plurality of reference samples. In deriving the one-dimensional array, the circuitry projects a value of at least one decoded pixel located on a first line onto a second line perpendicular to the first line, to derive at least one of the plurality of reference samples, and the smoothing is performed on the at least one decoded pixel projected onto the second line.