Video Coding Mode Selection Using Transformed Domain Error Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding algorithms face inefficiencies in bitrate/distortion performance due to high algorithmic complexity and processing time, particularly with a posteriori based methods, while a priori methods offer modest performance.
Innovation Solution
A method for selecting the best coding mode for a video block by estimating coding error using the sum of absolute values of coefficients after mathematical transformation and quantization, incorporating a predefined function of the quantization step, and weighting coding cost and header bits to reduce complexity and improve performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a posteriori based algorithm is used for coding mode decision, then bitrate/distortion performance is improved, but algorithmic complexity and processing time increase significantly
Solution Approach 1:
The patent applies preliminary action by estimating coding error using SATD (Sum of Absolute Transformed Differences) and quantized coefficient energy before actual macroblock reconstruction. This preliminary error estimation allows the algorithm to predict coding quality without performing the complete reconstruction process, thereby reducing computational complexity while maintaining accurate bitrate/distortion performance assessment
Solution Approach 2:
The patent extracts only the essential components needed for error estimation (SATD calculation and quantized coefficient energy) from the complete macroblock reconstruction process. By separating these estimation components from the full reconstruction pipeline, the algorithm achieves accurate mode decision without the computational burden of complete reconstruction, resolving the contradiction between performance accuracy and algorithmic complexity
2Measurement precision
If a posteriori based algorithm is used for coding mode decision, then bitrate/distortion performance is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary error estimation using SATD and quantized coefficient energy calculations before actual macroblock reconstruction. This preliminary assessment provides sufficient information for mode decision without requiring complete reconstruction, significantly reducing processing time while maintaining accurate bitrate/distortion performance evaluation
Solution Approach 2:
The patent skips the computationally intensive macroblock reconstruction step by using error estimation based on transformed domain coefficients. This allows the algorithm to rush through the mode decision process using simplified calculations that capture the essential quality metrics without performing the full reconstruction pipeline, thereby reducing processing time
3Measurement precision
If macroblock reconstruction is performed for each coding mode, then coding error estimation accuracy is improved, but algorithmic complexity increases
Solution Approach 1:
The patent extracts only the necessary error estimation components (SATD and quantized coefficient energy) from the complete macroblock reconstruction process. By taking out these specific measurement elements and performing estimation in the transformed domain, the algorithm achieves accurate coding error assessment without the complexity of full spatial domain reconstruction
Solution Approach 2:
The patent substitutes the mechanical reconstruction process (spatial domain operations) with mathematical operations in the transformed domain. By replacing the complex reconstruction mechanics with SATD calculation and energy-based coefficient analysis, the algorithm achieves equivalent or superior estimation accuracy with reduced computational complexity
Data Source
AI summary
The method comprises a coding of a source block according to a coding mode wherein the selection of the coding mode is carried out, for a source block, according to coding error estimations and source block coding cost estimations for the different coding modes tested, the coding error being estimated, for a coding mode, according to the sum of the absolute values of coefficients obtained after a mathematical transformation in the frequency domain of the residue block related to this coding mode and as a function of the sum of absolute values of quantized coefficients corresponding to this coding mode. An application is the decision of the coding mode for data using the MPEG 2 or MPEG4-AVC standard.