Motion Estimation Regions in Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High computational complexity and increased memory requirements in motion estimation for video encoding, particularly in High Efficiency Video Coding (HEVC), due to the need for separate motion vector determination for each prediction unit (PU), which hampers encoding efficiency and increases bitrate.
Innovation Solution
Divide the largest coding unit (LCU) into motion estimation regions (MERs) of equal size, performing first and second motion searches to determine quad-trees based on a cost function, thereby reducing the number of motion searches and simplifying the determination of motion vectors for PUs, and selecting the optimal quad-tree for encoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If separate motion vector determination is performed for each prediction unit (PU) in HEVC, then motion estimation accuracy is improved, but computational complexity and memory requirements increase
Solution Approach 1:
The patent divides the Largest Coding Unit (LCU) into multiple Motion Estimation Regions (MERs), each of which can be independently processed. This segmentation allows motion estimation to be performed at a coarser level (MER level) rather than at every PU level, reducing the total number of motion searches required while still maintaining adequate motion representation for each region. The MER structure enables hierarchical processing where motion vectors can be determined at multiple scales.
Solution Approach 2:
The patent merges multiple PUs within the same MER to share common motion estimation results. By combining the motion estimation process across multiple PUs at the MER level, the system reduces redundant computations that would otherwise occur if each PU were processed independently. This merging approach maintains motion accuracy for individual PUs while eliminating duplicate motion search operations.
2Loss of substance
If HEVC encoding is implemented to improve compression efficiency, then bitrate is reduced, but computational complexity increases
Solution Approach 1:
By segmenting the LCU into MERs with predetermined sizes, the patent creates a hierarchical structure that reduces the total number of motion estimation operations. Instead of performing full motion estimation at every PU level across the entire LCU, the system performs motion estimation at the MER level first, then selectively refines at the PU level only where necessary. This segmentation approach maintains HEVC's compression efficiency while reducing computational complexity.
Solution Approach 2:
The patent applies partial motion estimation by performing full motion searches only at the MER level and using simplified or selective motion estimation at the PU level. Rather than performing exhaustive motion estimation at every possible granularity level, the system performs partial motion estimation that is sufficient for the application requirements, thereby reducing computational complexity while maintaining adequate compression efficiency.
3Manufacturing precision
If multiple motion searches are performed to determine optimal quad-trees, then encoding accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary motion estimation at the MER level before finalizing PU-level motion vectors. By conducting initial motion searches at the coarser MER level, the system establishes baseline motion parameters that guide subsequent finer-grained PU-level estimation. This preliminary action reduces the search space for subsequent motion estimation steps, maintaining encoding accuracy while reducing overall processing time through hierarchical coarse-to-fine processing.
Data Source
AI summary
A method comprises dividing a largest coding unit (LCU) of a picture into a plurality of motion estimation regions (MERs) having size equal to or less than a predetermined size. For one or more MERs of the plurality of MERs, a number of first motion searches are performed for determining a first quad-tree based on a cost function associated with a first plurality of prediction units (PUs) of the one or more MERs. A number of second motion searches are performed for the LCU, for determining a second quad-tree, based on the cost function associated with a second plurality of PUs of the LCU. The first quad-tree or the second quad-tree is selected for performing encoding of the picture based on a comparison of a first cost of the first quad-tree with a second cost of the second quad-tree.