Coding Unit Division Predecision via Pixel Gradient Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High calculation complexity in determining the coding cost of coding units (CUs) in HEVC video coding leads to low coding efficiency and increased processing time.
Innovation Solution
A method that determines the coding cost of CUs by analyzing pixel gradient information to predecide on division, allowing for early termination of CU division and reducing the workload, using a server with modules for CU determination, gradient information calculation, predecision making, and cost determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the HEVC standard uses a flexible block structure with multiple CU sizes and tree-structure traversal for division, then the coding precision and adaptability are improved, but the device complexity and calculation complexity increase significantly
Solution Approach 1:
The patent performs preliminary actions by calculating pixel gradient information and making division predecisions before actual CU division and predictive coding. This early assessment of texture complexity allows the system to avoid unnecessary detailed processing for CUs that clearly require or do not require division, thus reducing overall calculation complexity while maintaining coding precision.
Solution Approach 2:
The patent applies partial action by selectively performing full predictive coding only on CUs that pass the gradient-based predecision filter. For CUs with simple texture characteristics, the system performs only the lighter gradient calculation and skips the expensive predictive coding step, achieving a balance between accuracy and complexity.
2Manufacturing precision
If the HEVC standard performs predictive coding and coding cost calculation for every CU in the tree structure, then the coding precision is improved, but the productivity and coding speed deteriorate
Solution Approach 1:
The patent performs preliminary gradient calculation and division predecision before the expensive predictive coding step. This preliminary assessment identifies CUs that can be safely skipped, allowing the system to maintain high coding precision for complex regions while dramatically improving overall coding speed by avoiding redundant calculations on simple regions.
Solution Approach 2:
The patent extracts and evaluates specific gradient features from each CU to make early division decisions. By taking out only the essential gradient information needed for predecision and using it to filter out CUs that don't require full processing, the system achieves both high precision where needed and high speed where possible.
3Reliability
If the HEVC standard performs complete traversal and recursion for CU division decision, then the reliability of division policy determination is improved, but the loss of time and processing efficiency worsen
Solution Approach 1:
The patent performs preliminary gradient-based division predecision for each CU before committing to full tree-structure traversal and recursion. This preliminary filter maintains reliability by using mathematically sound gradient analysis to make accurate early decisions, while simultaneously reducing processing time by avoiding complete traversal for CUs that can be decisively classified as simple or complex.
Data Source
Figure 1~2B
Figure 3A
Figure 3B
AI summary
This application discloses a method and an apparatus for determining a coding cost of a coding unit (CU). The method includes: determining a CU subject to predictive coding in an intra-frame prediction mode, determining pixel gradient information corresponding to the CU, the pixel gradient information including a pixel gradient variance, and the pixel gradient variance being a variance of image gradients of at least some pixels in the CU; making a division predecision on the CU according to the pixel gradient information; and determining a first coding cost as a second coding cost of the CU in a case that a result of the division predecision on the CU is negative, the second coding cost being used for determining a division policy of a coding tree unit (CTU) corresponding to the CU, and the first coding cost being a coding cost obtained by performing predictive coding on the CU by using a current size of the CU as a size of a prediction unit (PU). In technical solutions of this application, division of the CU is ended in advance, so that calculation workload may be reduced, coding complexity may be reduced, and a coding rate may be improved.