Image Coding Tree Unit Partitioning for Bandwidth and Cost Trade-offs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods struggle to independently determine coding tree unit size and maximum coding unit size, leading to increased implementation costs and reduced bandwidth efficiency due to the interdependence of these sizes.
Innovation Solution
A method that acquires coding tree unit information and signaling information on the maximum coding unit size, allowing for stepwise partitioning of the coding tree unit, thereby enabling independent determination of these sizes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If the coding tree unit size is increased to reduce line buffer bandwidth, then the implementation cost of the encoder and decoder increases due to larger maximum coding unit size
Solution Approach 1:
The patent divides the coding tree unit into multiple coding units using a quadtree structure, allowing the maximum coding unit size to be independently controlled from the coding tree unit size. This segmentation enables the system to use larger coding tree units for bandwidth efficiency while maintaining smaller maximum coding unit sizes for lower implementation cost.
Solution Approach 2:
The patent introduces dynamic partitioning where the maximum coding unit size can be adaptively adjusted through signaling information (maxCuDiffDepth) while the coding tree unit size remains fixed. This dynamic approach allows optimization of both bandwidth usage and implementation cost based on content characteristics.
2Device complexity
If the maximum coding unit size is decreased to reduce implementation cost, then the line buffer bandwidth increases
Solution Approach 1:
By segmenting the coding tree unit into multiple coding units with configurable maximum sizes, the system can process smaller blocks (reducing implementation cost) while using larger coding tree units (reducing line buffer bandwidth requirements through better memory utilization patterns).
Solution Approach 2:
The patent changes the parameter relationship by allowing maximum coding unit size to be independently signaled and adjusted separately from coding tree unit size, enabling optimization of both line buffer bandwidth and implementation cost through parameter tuning.
3Adaptability or versatility
If the coding tree unit size and maximum coding unit size are determined independently, then signaling information is required to transmit the size difference
Solution Approach 1:
The patent uses maxCuDiffDepth as an intermediary parameter that efficiently represents the relationship between coding tree unit size and maximum coding unit size. This single signaling value enables independent size determination while minimizing the amount of information that needs to be transmitted.
Solution Approach 2:
The patent changes from transmitting multiple size parameters to transmitting a single difference parameter (maxCuDiffDepth), reducing signaling overhead while maintaining the capability for independent size determination.
Data Source
AI summary
An image decoding method according to an embodiment of the present invention comprises the steps of: acquiring encoding tree unit information from an image stream; acquiring signaling information on a maximum coding unit size of a coding unit divided from the encoding tree unit; and dividing the encoding tree unit by steps according to the signaling information.