Video Encoding Tree Search for Motion Vector Bit Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
H.264/AVC video coding standard faces challenges in selecting optimal motion vectors due to increased complexity and side information requirements with multiple prediction modes, leading to suboptimal performance across various bit-rates.
Innovation Solution
An AND-OR tree-based encoding strategy is employed to structure coding options, incorporating residual coding into motion vector selection and using a tree-based representation to find a parameter set with minimum projected costs, which optimizes bit-allocation and improves prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple prediction modes are used to improve prediction accuracy, then prediction accuracy is improved, but the complexity of obtaining optimal coding parameters increases
Solution Approach 1:
The patent segments the complex parameter optimization problem into a structured AND-OR tree representation, where different prediction modes and their associated parameters are organized hierarchically. This segmentation allows systematic exploration of coding options without overwhelming complexity, as each node in the tree represents a manageable subset of parameters to optimize.
Solution Approach 2:
The patent performs preliminary action by constructing the complete AND-OR tree representation of all coding options before actual encoding. This pre-organization of coding strategies and their dependencies enables efficient navigation and selection during encoding, avoiding the need to re-evaluate complex interactions between prediction modes during the time-critical encoding process.
2Measurement precision
If more prediction modes are used, then prediction accuracy is improved, but the amount of side information required increases
Solution Approach 1:
The patent applies dynamics by making the selection of prediction modes and their parameters adaptive to the local characteristics of each video block. The AND-OR tree structure enables dynamic selection of the most appropriate coding strategy for each block based on its specific properties, rather than uniformly applying all prediction modes throughout the video sequence. This dynamic adaptation reduces the average side information requirement while maintaining high prediction accuracy where needed.
3Ease of manufacture
If Lagrangian optimization is used for motion estimation, then a practical solution is achieved, but optimal motion vector selection across all bit-rates is not obtained
Solution Approach 1:
The patent changes the parameter optimization approach from fixed Lagrangian multipliers to a bit-rate adaptive selection process using the AND-OR tree. By organizing coding options in a tree structure that explicitly represents dependencies between parameters and bit-rate constraints, the system can select motion vectors and prediction modes that are optimal for each specific bit-rate target, rather than relying on a single Lagrangian cost function that may not capture all bit-rate scenarios.
4Measurement precision
If residual coding is incorporated into motion vector selection, then optimal performance at specific bit-rates is achieved, but the complexity of the encoding process increases
Solution Approach 1:
The patent merges the motion vector selection process with residual coding evaluation within the unified AND-OR tree framework. Instead of treating these as separate sequential steps, the tree structure integrates both considerations at each node, allowing the encoder to jointly optimize motion vectors and residual coding parameters. This merging reduces overall complexity by eliminating redundant evaluations and enabling coordinated optimization of both aspects simultaneously.
Data Source
AI summary
An encoding system for a data set, particularly for a video data set is disclosed in the present invention. The encoding system includes a coding strategies represented as a tree-based structure, a cost projection system for each tree node, and a parameter searching algorithm for coding strategies to encode a data set. When the encoder parameters are obtained, the coding strategies and rate-allocation have better performance comparing to the related art.