Image Processing Apparatus Motion Vector Alignment Skip Block Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image encoding methods, such as H.264 and HEVC, face challenges in reducing calculation costs for motion vector search and aligning motion vectors, particularly when using binary or N-value images, which can lead to lower encoding efficiency and difficulty in employing skip blocks.
Innovation Solution
An image processing apparatus that determines whether to use a reference vector as a motion vector for a block based on the encoding result of the previous frame, allowing for improved alignment of motion vectors and increased use of skip blocks by employing a reference vector or N-value image search vector depending on the encoding mode.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If motion vector search is performed using binary image, then calculation cost is reduced, but motion vector alignment becomes difficult and skip block cannot be performed
Solution Approach 1:
The patent dynamically switches between binary image search and multi-value image search based on the block type and encoding mode. For skip blocks and certain inter-blocks, binary image search is used to reduce calculation cost. For other blocks where motion vector alignment is critical, multi-value image search is performed to ensure accurate motion vector acquisition, thereby resolving the contradiction between calculation cost reduction and motion vector alignment ease.
Solution Approach 2:
The patent applies different search methods to different blocks based on their specific characteristics. Skip blocks and certain inter-blocks use binary image search for efficiency, while other blocks use multi-value image search for accuracy. This local differentiation allows the system to optimize both calculation cost and motion vector alignment performance according to local block requirements.
2Measurement precision
If full search is performed for motion vector, then motion vector accuracy is improved, but calculation cost becomes huge
Solution Approach 1:
The patent uses binary image as a simplified copy of the original multi-value image for motion vector search. The binary image retains the essential structure and edge information needed for motion vector acquisition while removing redundant pixel value information, thereby reducing calculation cost while maintaining sufficient motion vector accuracy for most blocks.
Solution Approach 2:
The patent performs partial search using binary image for blocks where high accuracy is not critical, and reserves full multi-value image search for blocks where motion vector accuracy is essential. This partial action approach allows the system to reduce overall calculation cost while maintaining sufficient precision for critical blocks.
3Productivity
If skip block is used to reduce data amount, then encoding efficiency is improved, but motion vector alignment requirement becomes more strict
Solution Approach 1:
The patent dynamically determines whether to perform binary image search or multi-value image search based on the block type. For skip blocks, binary image search is used to reduce calculation cost, while for blocks requiring precise motion vector alignment, multi-value image search is performed. This dynamic adaptation allows the system to maintain encoding efficiency while ensuring sufficient motion vector alignment precision.
Data Source
AI summary
A variable-length encoding unit in an image processing apparatus encodes a frame constituting a moving image, and a selection unit determines whether to employ a reference vector calculated from a motion vector corresponding to a block other than a first block among a plurality of blocks in a second frame subsequent to a first frame constituting the moving image as the motion vector corresponding to the first block, based on an encoding result of the first frame.