Video Encoding Quantization Matrix Set Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing H.265/HEVC video coding standard is inefficient in reducing the amount of codes required, leading to increased bitstream occupation and difficulties in image compression, particularly when user-specified quantization matrices are used, which often results in higher image quality but increased code amounts.
Innovation Solution
An encoding method that selects a quantization matrix set on a per-block basis, using a combination of arbitrarily specified and predefined quantization matrix sets, where each set has unique quantization values, allowing for efficient encoding and decoding by reducing the frequency of custom matrix usage and optimizing code allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If user-specified quantization matrices are used to improve image quality, then subjective image quality is increased, but the amount of codes increases
Solution Approach 1:
The patent pre-defines multiple quantization matrix sets with different characteristics (first quantization matrix set, second quantization matrix set, third quantization matrix set) before encoding. These pre-defined sets cover different quantization strengths and frequency component handling strategies, allowing the encoder to select the most appropriate set without needing to transmit custom matrix data, thereby reducing code amount while maintaining image quality
Solution Approach 2:
The patent changes the parameter of quantization matrix selection by introducing multiple pre-defined quantization matrix sets with different characteristics. Instead of using a single fixed quantization matrix or requiring user-specified custom matrices, the system varies the quantization parameters by selecting from multiple pre-configured sets, each optimized for different image characteristics and compression requirements
2Productivity
If multiple quantization matrix sets are defined and selected based on block features, then code allocation is optimized, but the complexity of setting and selecting increases
Solution Approach 1:
The patent applies different quantization matrix sets to different image blocks based on their local characteristics. Each block is analyzed for its features (such as frequency content, texture, and importance), and the most suitable quantization matrix set is selected for that specific block. This localized approach optimizes compression efficiency for each region without requiring complex global optimization
Solution Approach 2:
Multiple quantization matrix sets are pre-defined and prepared before the encoding process begins. These sets include various configurations optimized for different image characteristics. During encoding, the system only needs to select from these pre-prepared sets based on block features, avoiding the complexity of creating and optimizing matrices in real-time
Data Source
AI summary
An encoding method which allows reduction in the amount of codes is provided. The encoding method includes: setting step S11 for setting a quantization matrix set; quantization step S12 for performing quantization on a target block using the selected quantization matrix; and encoding step S13 for encoding, in a mutually associated manner, the target block which has been subjected to the quantization and identification information for identifying the quantization matrix set which has been set. In setting step S11, a quantization matrix set selected from a plurality of quantization matrix sets is set as the quantization matrix set to be used to perform the quantization on the target block, the plurality of quantization matrix sets including a custom quantization matrix set which is arbitrarily specified and a plurality of defined quantization matrix sets which have been respectively defined in advance.