Sign Prediction in Video Coding Using Subsampled Block Borders
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently reducing redundancy in video signals, particularly in sign prediction of transform coefficients, which affects compression efficiency and computational complexity.
Innovation Solution
The proposed solution involves processing circuitry that calculates the cost of sign prediction hypotheses for a subset of transform coefficients based on samples in the topmost or leftmost rows/columns of a block, selecting the hypothesis with minimal cost, and performing sign prediction accordingly, with the number of samples being a constant number such as 4 or 8, and determining this number based on block dimensions or thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sign prediction is performed on all transform coefficients using all border samples, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent segments the transform block into multiple regions (e.g., top-left, top-right, bottom-left, bottom-right regions) and applies different sign prediction hypotheses to each region. This allows the encoder to test multiple hypotheses only on a subset of coefficients rather than all coefficients, reducing computational complexity while maintaining prediction accuracy for the most significant coefficients.
Solution Approach 2:
The patent extracts and processes only the most significant transform coefficients (e.g., DC coefficient and selected AC coefficients) using sign prediction, while bypassing less significant coefficients. This extraction approach focuses computational resources on the coefficients that contribute most to reconstruction quality, thereby reducing overall computational complexity.
2Productivity
If more sign prediction hypotheses are tested, then compression efficiency is improved, but processing time increases
Solution Approach 1:
The patent tests a limited number of sign prediction hypotheses (e.g., 2-4 hypotheses) rather than exhaustively testing all possible hypotheses. This partial action approach achieves sufficient compression efficiency by considering the most likely sign patterns while avoiding the excessive processing time that would result from testing all possible hypotheses.
Solution Approach 2:
The patent applies different numbers of sign prediction hypotheses to different regions of the transform block based on local characteristics. For example, regions with higher energy coefficients may receive more hypotheses while lower energy regions receive fewer hypotheses, optimizing the balance between compression efficiency and processing time.
Data Source
AI summary
An apparatus includes processing circuitry that calculates, for each of a plurality of sign prediction hypotheses, a cost of the respective sign prediction hypothesis by performing sign prediction on a subset of a plurality of transform coefficients based on a subset of samples in at least one of a topmost row or a leftmost column of a current block. A number of samples in the subset of samples is a constant number and less than a total number of samples in the at least one of the topmost row or the leftmost column of the current block. The processing circuitry selects one of the plurality of sign prediction hypotheses with a minimal cost among the costs of the plurality of sign prediction hypotheses. Further, the processing circuitry coded the current block according to the selected sign prediction hypothesis.