Low-Latency Illumination Compensation in 3D Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing three-dimensional (3D) and multi-view video coding techniques face latency issues in illumination compensation (IC) and Depth Lookup Table (DLT) design, leading to inefficiencies in encoding and decoding processes.
Innovation Solution
The method determines whether to enable or disable IC and DLT based on statistical conditions related to selected reference pictures or slices, excluding pixel value information, allowing for low-latency decision-making and reducing processing latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real-time pixel value analysis is performed to determine IC and DLT parameters, then coding accuracy is improved, but processing latency increases
Solution Approach 1:
The patent applies preliminary action by determining IC and DLT parameters for future pictures using reference pictures from the past. The encoder uses statistics from previously decoded reference pictures to predict parameters for current pictures, eliminating the need for real-time analysis. This is achieved by selecting reference pictures with smaller temporal IDs and using their statistics to determine parameters for pictures with larger temporal IDs, thus performing the analysis work in advance.
Solution Approach 2:
The patent uses copying by replicating parameter determination from reference pictures to current pictures. Instead of analyzing pixel values in real-time, the system copies parameter values from previously decoded reference pictures that have similar characteristics. This allows the decoder to reuse parameters without performing new analysis, significantly reducing processing latency while maintaining coding accuracy.
2Measurement precision
If statistics from multiple reference pictures are analyzed to determine IC parameters, then parameter accuracy is improved, but encoding complexity increases
Solution Approach 1:
The patent applies local quality by selecting specific reference pictures based on their temporal ID relationships with the current picture. Instead of uniformly processing all reference pictures, the system identifies and processes only those reference pictures with smaller temporal IDs that are most relevant to the current picture. This selective approach reduces the number of statistics that need to be analyzed while maintaining parameter accuracy.
Solution Approach 2:
The patent changes parameters by using temporal ID comparisons to dynamically select which reference pictures to use for parameter determination. The system adjusts the selection criteria based on the temporal relationship between reference and current pictures, optimizing the balance between parameter accuracy and encoding complexity by processing only the necessary subset of reference pictures.
3Reliability
If IC is enabled for all CUs to ensure consistent quality, then video quality is maintained, but bitrate increases
Solution Approach 1:
The patent applies partial action by enabling IC only for CUs where it is actually needed, rather than uniformly for all CUs. The system uses the determined IC parameters to selectively apply illumination compensation only to those coding units that benefit from it, based on the statistical analysis of reference pictures. This selective approach maintains video quality where necessary while reducing overall bitrate compared to universal IC application.
Data Source
AI summary
A method and apparatus for low-latency illumination compensation in a three-dimensional (3D) and multi-view coding system are disclosed. According to the present invention, the encoder determines whether to enable or disable the illumination compensation for the current picture or slice based on a condition related to statistic associated with a selected reference picture or slice respectively, or related to high-level coding information associated with the current picture or slice respectively. The high-level coding information associated with the current picture or slice excludes any information related to pixel values of the current picture or slice respectively. The illumination compensation is them applied according to the decision made by the encoder. A similar low-latency method is also applied for depth lookup table (DLT) based coding.